AI Digital Agency — Idea Snapshot

All strategy data captured for the AI Digital Agency business idea (Sessions 1–3 of co-founder workstream).

Generated 2026-05-09 · 9 strategy documents · Brand name TBD · Bootstrap-mode launch · France-first, ICP 1 (Wellness/Beauty) bullseye

Business Case

Wedge, why now, value proposition, locked working assumptions.

Source: ~/claude-sync/business/products/ai-agency/business-case.md

AI Digital Agency — Business Case

Captured: 2026-05-07 (Alfred + Boss, co-founder kickoff session) Status: Locked. Inputs to the next artifacts (lean plan, brand, PRD, etc.). Companion docs: ../bundle-presence-locale.md (SUPERSEDED, historical), ../../strategy/decisions.md → 2026-05-07 entries.

Session 2 amendments (2026-05-07 evening):

  • Wedge gains a fourth dimension: 24/7 customer support. SMB owner can reach our AI customer support agent any time, any day. Traditional agencies don't. The wedge is now price + speed + quality + always-on.
  • Stated assumption #1 flipped: paid API from day 1, not Claude Max sub. Per-customer token spend is now a real cost line the lean plan must size and cover.
  • 24/7 customer support is table-stakes inclusion, not a separate service. Confirmed by Boss. Attached automatically to whatever the customer bought (bundle or a la carte). Scope: regular customer support — billing questions, troubleshooting, and capped small content edits per month (e.g., swap a photo, change a price, edit a text block). The cap number is a pricing-tier lever, set at lean-plan / pricing time.

One-pager summary

We deliver the same outcomes as a traditional digital agency — working website, Google Business Profile, review system, social content, SEO/GEO, digital strategy, AI chatbot — to non-technical small-business owners, at one-tenth the price, with same-day delivery, and 24/7 customer support (SMB owner can reach our AI agent any time, any day). Made possible by AI replacing the agency labor that creates the cost. Customers come because we're cheaper, faster, always-on, and the quality matches. The "AI agency" framing is brand language, not the wedge.

1. The problem

Non-technical small-business owners want the outcomes of professional digital presence — more walk-ins, more callbacks, more bookings, real reviews, a credible online face — but agency prices that deliver this (€1,500–5,000 setup + €300–500/month retainer) are out of reach for most SMBs. Self-serve tools (Wix, Lovable, Canva, Hootsuite, Buffer) are affordable but require operating a tool the buyer doesn't want to operate. Result: SMBs run with no real digital presence, or buy a one-shot €350 site that goes stale within months — and never come back.

The pain is at the outcome level ("I want more customers"), not the artifact level ("I want a website"). The buyer wants the result; they don't want to build it themselves and they can't afford an agency to build it for them.

2. The buyer (ICP)

Size: Micro to small SMBs, 1–30 employees. Decision-maker: Owner-operator (same person who runs the business, signs the cheque, feels the pain). Filter: Verticals where Google search / Google Maps / reviews drive real customer acquisition.

Strong-fit verticals for V1 cold outreach:

  • Trades — plumbers, electricians, AC repair, locksmiths, painters, garages, contractors
  • Search-driven professional services — immigration lawyers, divorce lawyers, traffic-ticket lawyers, small-business accountants, notaires for search-driven niches
  • Local discovery services — vet clinics, opticians, fitness studios, dental / aesthetic / GP clinics, salons, spas, restaurants, cafés

Excluded from V1:

  • Hospitality (booking systems require transactional features — out of V1 scope)
  • Pure referral-only businesses (corporate law, M&A, B2B strategy consulting — Google isn't their funnel)
  • Anything requiring transactional site features (e-commerce, real-time booking, member portals)

Offer structure: Bundle + a la carte. Customers pick what they actually need. The AI onboarding agent at intake recommends bundle vs. a la carte based on the customer's stated goals.

V1 site constraint: Brochure sites only — services, pricing, contact info, hours, gallery, embedded GBP/Maps, click-to-call. No e-commerce, no booking systems, no member portals, no custom backend integrations.

3. The wedge

Same agency outputs (site + GBP + review system + social content + SEO/GEO + strategy + AI chatbot), at one-tenth the price (€19–99/month vs. €300–500/month retainer), with same-day or next-day delivery (vs. weeks at a traditional agency), and 24/7 customer support — the SMB owner can reach our AI customer support agent any time, any day, in FR / AR / EN. Traditional agencies close at 6pm and ignore weekends; we don't, because the support agent is AI. Made possible by AI replacing the agency labor that creates the cost.

Why us:

  • Pre-built infrastructure — Content Factory, Claude Code social MCPs, GA4/GSC/Vibe Prospecting MCPs, ReadyGo template patterns. Roughly 6–12 months of head start over anyone trying to build this from zero.
  • Solo + AI cost structure — no payroll, no office overhead, marginal cost per customer is mostly AI tokens. Price points incumbent agencies can't profitably match.
  • Existing agency operating experience (ReadyGo) — direct knowledge of SMB sales cycle, real templates, real client patterns to learn from.
  • Three working AI agents already in production — unusual depth in agent orchestration vs. typical SMB-services founder.

Why now:

  • LLM costs have crashed (Claude Sonnet ~$3/million input tokens). Per-customer content generation in FR/AR/EN is now economically viable in a way it wasn't 18 months ago.
  • Self-serve AI tool fatigue is real — Lovable / v0 / Bolt have proven the technology works AND that non-technical buyers don't actually adopt these tools. The category gap (services vs. tools) is now visible.
  • SMB digital adoption is accelerating in many markets — buyers are searchable, reviewable, reachable in a way they weren't 5 years ago.
  • Window before competitors close the gap: roughly 18–36 months. Established agencies retool with AI; big platforms (Wix / Squarespace / Shopify) launch fully-managed AI service tiers; multiple "AI agency" startups appear globally.

Defensibility: This is a window play, not a permanent moat. Defensibility builds later through customer base, brand, regional category leadership, operational learning, and proprietary delivery infrastructure — not from day-one moats. The first 12 months are about using the head start to lock down a regional position before competitors arrive.

4. Market

Buyer universe: Globally there are tens of millions of micro and small businesses (1–30 employees) that fit the buyer profile above. Even narrowing to a single language or single geography, we're talking about hundreds of thousands to millions of potential customers. The market is not the constraint.

Real ceiling for V1: What a solo founder can serve, not what the market contains.

Rough capacity:

  • Per-customer onboarding: ~1–1.5 hours of founder time
  • Ongoing support: ~10 minutes per active customer per month
  • Cold outreach + sales calls: significant ongoing time allocation

A solo founder running this stack can probably handle 500–1,500 active paying customers before the math breaks — at which point we'd need a hire, more aggressive automation, or both.

Revenue ceiling at solo-founder capacity, blended €40/month average across bundle + a la carte:

  • 100 customers ≈ €48k ARR
  • 300 customers ≈ €144k ARR
  • 500 customers ≈ €240k ARR
  • 1,000 customers ≈ €480k ARR
  • 1,500 customers ≈ €720k ARR

Honest claim signed:

The buyer universe is large enough not to be a constraint. The real ceiling is what a solo founder can serve — roughly 500–1,500 active customers — which corresponds to €200–700k yearly recurring revenue at our pricing. That's a real, profitable solo-founder business at every step from 100 customers up.

Note: launch market choice is a separate decision. This business case is market-agnostic. The specific country/region we launch in gets locked in the GTM session, not here.

5. The bet

AI in 2026 is good enough that a properly orchestrated stack delivers output the SMB buyer rates as agency-grade or better — at one-tenth the cost a traditional agency charges to deliver the same.

That single claim is what the entire business stands on. Everything else is downstream.

Why we believe it: Founder has operated the stack and verified the output bar. Sometimes the output beats human agency work. That's not investor-grade proof but it's founder-level conviction backed by direct evidence. Combined with the LLM cost trajectory, this is now economically real.

Why competitors don't believe it: Most agencies, even ones using AI internally, treat AI as a productivity tool for human workers — not as the workforce itself. They still price as if humans are doing the work because in their setup, humans are. They believe agency-grade quality requires human craftspeople for design judgment, copy nuance, and client handling. We bet they're holding onto that belief out of habit, not because the quality gap is still real.

Falsification signals (we know in 90 days from first paying customer):

  • First 5 paying customers consistently rate output below "what I'd expect from an agency"
  • Churn rate above 30% in first 6 months
  • Refund requests citing quality, not other reasons
  • Negative word-of-mouth in customers' local circles

If those signals appear, the bet failed and we pivot fast. Cheap to test, fast to detect.

If we're right: the cost structure carries the business. Same outputs as agencies, ten times cheaper, days instead of weeks — that's a sales conversation that wins on its own. Customers come for the price-quality-speed combo. The "AI agency" framing becomes useful brand language but isn't load-bearing.

Services (V1)

7 services, each sellable a la carte or as a bundle.

1. Website

Static informational website (services + pricing + contact info, no transactional features). Built from a vertical-specific template, customized via AI from the customer's intake answers. Hosted on our infrastructure. Same-day delivery. Outcome: Customer has a credible, searchable web presence with their own domain.

2. GBP (Google Business Profile)

For customers with an existing GBP → AI optimizes via MCP (fixes missing fields, normalizes inconsistent info, adds photos, hours, services, categories, posts updates). For customers without → human-assisted setup + verification (Google requires postcard / phone / video verification of business ownership, AI can't fake that), then AI takes over for ongoing optimization. Includes monthly health check + auto-posted updates / offers / events. Outcome: Customer is findable on Google Maps and Google search, with a fresh and complete business profile that ranks well locally.

3. Review Engine

Automated review request triggers (SMS / email / WhatsApp / plexi QR or sticker QR) after each completed service. Real-time monitoring of incoming reviews. Negative-review alerts to owner. AI-drafted response suggestions in FR / AR / EN. Monthly velocity report. Outcome: Customer's Google review count grows consistently from ~5% ask-rate to 60–80%, lifting local ranking and conversion.

4. Content (social media)

AI-generated social posts on Facebook + Instagram (LinkedIn for B2B niches when unblocked) on a fixed cadence (e.g., 4 / 16 / 30+ posts per month depending on tier). Auto-publishing on schedule. AI auto-reply on comments and DMs with vertical-tuned tone. Owner approves before posts go out (or full auto at higher tiers). Outcome: Customer maintains an active, on-brand social presence without lifting a finger.

5. SEO/GEO (full-service search optimization)

Keyword research + tracking. On-page optimization (titles, meta, headers, schema markup). Technical SEO (site speed, mobile, sitemaps, crawlability). Local SEO (NAP citations, local directories). Monthly blog content targeting tracked keywords. GEO layer: AI-visibility audit + schema/structured-data optimization + brand mention strategy so the customer gets cited in ChatGPT / Claude / Perplexity / Google AI Overviews when their service is asked about. Outcome: Customer ranks higher on Google for their target searches AND gets cited by AI assistants for their vertical/locale.

6. Strategy

At onboarding: AI-generated personalized digital strategy doc covering positioning, channel mix, content angles, KPIs, and 3-month action plan, based on the customer's intake answers and competitor scan. Refreshed every 6 months. Outcome: Customer has a clear written plan for what to do digitally and a recurring health check on whether it's working.

7. AI Chatbot

A simple chat widget on the customer's site that answers visitor questions (hours, services, pricing, area served, FAQs) and routes inquiries to the contact form. Trained on the customer's intake data. Outcome: Lost leads recovered — visitor questions get answered when the SMB owner isn't there.

Stated assumptions

These were flagged at the start of the session and run through the whole business case. Surfaced again here so they're not buried.

1. LLM runtime budget = paid API from day 1. UPDATED 2026-05-07 (Session 2 kickoff): Boss called this — Claude Max sub is not the production runtime. We go API at launch. Per-customer token spend becomes a real cost line in the unit economics, not absorbed by a flat sub. Boss commits to sizing the API budget at launch time; the lean plan must produce a per-customer monthly token estimate that pricing covers with margin.

2. Split-attention build, not full-time founder. Confirmed. Founder is also running existing ReadyGo work and managing parallel obligations. The 60–90 day "ship V1" timeline is part-time work on top of existing pipeline, not clean-room dedicated time. Holds until proof-of-concept is validated and cash flow is stable. Roadmap session needs to reflect this.

3. Founder-level conviction on AI quality. Confirmed — but with an honest caveat: the AI-Agency framing of this idea is 2 hours old at lock time. Founder has operated the stack and judged output to be at agency-grade or better. Not customer-validated yet. The bet (Section 5) makes this explicit. Watch for false certainty as we build the lean plan — if the math reveals something unexpected, we re-open the bet rather than defend it.

4. Custom design tier as a paid upgrade. Confirmed and tested. For verticals or customers who demand human-designed visuals, we offer a paid upgrade where a human designer (founder or sub-contracted) does the design work. Pricing must reflect real founder-time cost (€499–999 setup minimum on top of monthly) so it doesn't become the default path. Not a V1 must-have but reserved as an option.

Open items carried into next artifacts

These were raised in the kickoff session but belong in subsequent artifacts:

  • Pricing model + numbers — bundle tiers, a la carte per-service prices, custom design upgrade pricing. → Lean business plan session (after launch market is locked, since pricing depends on local price perception).
  • Launch market choice — which country/region we launch in first (Q11 in open-questions.md). → GTM plan session.
  • Brand container — under ReadyGo, under Digital Fanatic Ltd, or new brand entirely (cross-link to Q10). → Brand session.
  • Working name — current placeholder is "AI Digital Agency"; real brand name TBD. → Brand session.
  • AI onboarding agent V1 build spec — intake flow, quote logic, technical implementation. → PRD session.
  • Cold outreach playbook — channel mix, scripts per vertical, conversion mechanics. → GTM session.
  • Service-by-service unit economics — cost-to-deliver per service (founder hours + AI tokens + tools). → Lean business plan session.
  • Full operational scope per service — every flow, screen, automation, customer touchpoint. → PRD session.

Next artifact

Lean business plan / unit economics.

Inputs from this business case: the 7 services, the buyer profile, the wedge, the capacity ceiling, the stated assumptions.

Goal of next session: cost-to-deliver per service, candidate pricing structures (bundle tiers + a la carte), 12-month cash projection scenarios, breakeven math.

Positioning

Category-king strategy: "Site Vitrine" + "Visibility Starter Pack" — discovery anchors and naming conventions.

Source: ~/claude-sync/business/products/ai-agency/positioning.md

AI Digital Agency — Positioning

Locked Session 3, 2026-05-08. Foundation document for ALL marketing copy, ad creative, onboarding agent voice, content topics, FAQ wording, and brand language. Every piece of customer-facing language must trace back to this doc.

The strategic move

Category-king positioning. We're not just selling a service — we're claiming mental category ownership for two trigger states in our customer's head.

Like Slack owns "workplace messaging" and Stripe owns "payments for developers," we want to own:

Customer trigger state Our category label What we sell
"I need a website for my business" Site Vitrine Standalone Website service
"I'm starting a business / need visibility online" Visibility Starter Pack Essential bundle (Site + GBP + Reviews + Content + SEO)

Defaults win commerce. Customer doesn't shop "websites for SMBs" — they think "je vais prendre un Site Vitrine" and that means us. Same with the bundle: "je veux le Visibility Starter Pack".

Without these labels, we're "another digital service competing on features." With them, we're the default mental shortcut.

Discovery anchors (3-attribute hook)

The customer-facing acquisition message anchors on three of the four wedge dimensions:

  1. Rapide — Same-day delivery (vs agency 4-12 weeks)
  2. Abordable — 10× moins cher qu'une agence (NOT "cheap" — comparative undercutting, not bottom-tier signal)
  3. Qualité agence — Output quality of a junior agency staffer at scale

The 4th wedge dimension (24/7 support) lives downstream as a retention/satisfaction lever, not as discovery. It's not what closes the first sale; it's what justifies the subscription.

Important — proof not claim: "Qualité agence" must always be shown, never said. Demo galleries, before/afters, customer testimonials, real customer logos. Marketing copy never lazy-claims "haute qualité" — proof carries the assertion.

Full 4-dimension wedge (used in product/retention messaging)

For deeper copy, retention emails, customer success, and long-form content:

  1. Rapide
  2. Abordable (10× moins cher qu'une agence)
  3. Qualité agence
  4. Support 24/7

Use the 3-attribute hook for ad creative, headlines, landing pages above-the-fold. Use the 4-dim wedge in pricing pages, onboarding, FAQ, value-justification copy.

Naming conventions (lock these or the positioning never sticks)

Coined terms only land through zealous repetition. Use these exact labels with these exact casings, in every customer-facing surface:

Term When to use Variations to AVOID
Site Vitrine Standalone Website product. Always capitalized when used as our category name. "the website service," "notre service de site web," "la formule site"
Visibility Starter Pack Essential bundle. English-in-French is fine — works in French B2B tech. "the bundle," "la formule complète," "le pack" without "Visibility Starter"
Visibility Starter Pack Pro Auto-mode upgrade (was "+€19 add-on") —
Visibility Starter Pack Founders First-50 pricing (€59/mo) —
Site Vitrine Sur-Mesure Custom-design (€599 setup) tier "premium design," "version custom"

The rule: if you find yourself describing what we sell without using one of these labels, the copy is wrong.

Headline frameworks

Tagline candidates (not yet locked, all use category labels):

  • "Le réflexe Site Vitrine."
  • "Le Visibility Starter Pack pour SMB français."
  • "Site Vitrine. Rapide. Abordable. Qualité agence."

Above-the-fold landing page formulas:

  • "Le Site Vitrine de référence pour les SMB français — livré en 24h, 10× moins cher qu'une agence."
  • "Visibility Starter Pack: ton site, ton Google Business Profile, et ton contenu — tout en un."

Ad creative formulas:

  • "Tu cherches un Site Vitrine? Voici le réflexe."
  • "Le pack visibilité que les SMB attendent. €79/mois."
  • "Visibility Starter Pack — site + GBP + contenu + SEO. Livré ce soir."

Where this propagates (every surface)

Surface How positioning shows up
Reverse-pitch demos (Strategy A) "Voici TON Site Vitrine, déjà fait en 4 heures"
Partnership ad creative (Strategy C) "Le Site Vitrine que vos clients cherchaient" / "Le Visibility Starter Pack à recommander"
Onboarding agent opening "Tu cherches un Site Vitrine, ou le Visibility Starter Pack complet?"
Pricing page tiers Labels under each tier explicitly use "Site Vitrine," "Visibility Starter Pack," etc.
Content series headlines "5 erreurs à éviter sur ton Site Vitrine" / "Pourquoi le Visibility Starter Pack remplace ton agence"
FAQ headers "Pourquoi notre Site Vitrine est différent?" / "Que comprend le Visibility Starter Pack?"
Email subject lines "Ton Site Vitrine est prêt."
GBP description (our own) Lead with "Site Vitrine et Visibility Starter Pack pour SMB français"
Customer success communications Reinforce the label every interaction
Press / partnership pitches Lead with the category labels — make journalists/partners use our terms

Risks & failure modes

  1. Drift: team uses "the bundle" or "our website service" instead of the locked labels. Fix: copy reviews check label compliance every time.
  2. Anchor erosion: "Cheap" creeps back into copy. Always use "abordable" or "10× moins cher" — never "pas cher" or "cheap."
  3. Quality-claim laziness: "Haute qualité" or "qualité supérieure" written without proof. Always demonstrate visually, never assert verbally.
  4. Multi-language inconsistency: if/when we expand Tunisia (Arabic) or English markets, the category labels need direct translation strategy. For now: French only, "Visibility Starter Pack" stays English.

Cross-doc dependencies

  • business-case.md — wedge claims must stay consistent with positioning
  • lean-plan.md — pricing math; "10× moins cher" must hold mathematically against agency benchmarks
  • customer-profiles.md + customer-personas.md — the language used by personas should resonate with the labels
  • scope-boundaries.md — what we promise; positioning must not over-promise beyond delivery boundaries
  • marketing-strategies.md — all strategies execute the positioning
  • partnerships-strategy.md — partner-facing materials use the same labels

Status

Concept locked Session 3, 2026-05-08. Labels and discovery anchor finalized. Headline copy is directional, not final — gets refined when copywriting begins. Brand name (the agency itself, not the product categories) still pending; brand session is a separate milestone before launch.

Scope Boundaries

What we deliver and what we do not. Per-service tables, exclusions, customer-fit "not us".

Source: ~/claude-sync/business/products/ai-agency/scope-boundaries.md

AI Digital Agency — Scope Boundaries

Locked Session 3, 2026-05-08. This is the canonical reference for what we deliver and what we don't. Marketing copy, sales pages, FAQ, and onboarding-agent answers all cite back here. Anything we promise outside this table = trust debt the moment a customer signs up.

Per-service delivery table

Service What we deliver What we don't deliver
Website (€29/mo) Mockup-gallery template populated from intake; same-day live site; mobile responsive; basic SEO foundation; hosting included; cosmetic-edit support Bespoke design (paid add-on €599 setup); e-commerce checkout (Woo/Stripe); custom integrations (CRM, booking, calendars); multi-language native handling; advanced animation / scrollytelling; portable code ownership on exit
GBP + Reviews (€29/mo) GBP optimization (description, categories, hours, services, photos, posts); review solicitation engine via SMS / WhatsApp deep-link (sends from owner's phone, €0/send); Q&A seeding & monitoring; performance reporting Manufactured / fake reviews; verification done for the customer (we coach via written + video guide); algorithm / Maps-ranking guarantees; photo creation (we use customer-uploaded with format guidance)
Content / Social (€49/mo) 16 posts/mo on Instagram + Facebook; image + caption + hashtags; brand-voice consistency derived from intake; scheduling automation via Postiz TikTok / LinkedIn (V2); video production; DM / community management; paid ad campaign management; trending-audio or cultural-moment posts; manufactured engagement
SEO Basic (€39/mo — bundle-only with Website) Keyword research; on-page optimization; basic technical SEO; monthly content optimization on the existing site Backlink campaigns; digital PR; citation network building; competitor displacement work; ranking position guarantees; visible results before 3–6 months; SEO Advanced (deferred V2, €99/mo standalone, gated on hire)
Strategy (€15/mo add-on) 6-month locked strategy PDF + €10 on-demand refresh; marketing/digital strategy recommendations from intake patterns Consulting calls or live debate cycles; vertical-deep bespoke research; financial modeling; organizational / executive strategy; iterative collaboration
24/7 Support (included) AI-handled tier 1, always-on, for edits / questions / troubleshooting; founder escalation during business hours; cosmetic edits within a defined monthly limit Redesigns; new feature builds; business consulting; custom code work; crisis-communication management; unlimited cosmetic edits

Categorically outside our scope (V1 and likely beyond)

What Why we don't do it
Logo design, brand identity, photography, video production Not a creative agency; human craft AI can't yet replace at our price point
Google / Meta / TikTok ads management Different discipline with different unit economics; conflicts with our subscription wedge
Email marketing / CRM / marketing automation strategy Outside the templated subscription scope
Custom development (apps, software, bespoke integrations) Not a dev agency
Consulting hours / strategy debate sessions Our model is decisive AI-driven delivery, not collaborative consulting

Deferred to V2

What Why deferred
AI Chatbot for customer's site Out of scope for V1; not core to wedge
GEO (generative engine optimization for ChatGPT / Perplexity / etc.) Market still nascent; revisit when methodology stabilizes
TikTok + LinkedIn social posting Complexity vs. focus; V1 anchors on Instagram + Facebook
AI image generation Cost / quality threshold for client-facing output not yet locked
SEO Advanced (standalone, €99/mo) Founder-time bottleneck — 80 min/customer/mo × 100 customers = 133 hr/mo. Gated on first hire.

Customer-fit "not us"

Customer profile Where they should go instead
Enterprise (SSO, compliance audits, procurement workflows) Traditional agencies or enterprise SaaS
Multi-location complex businesses Wait for V2 multi-location tier; or traditional agency for now
Heavily regulated verticals (medical, legal, finance) without compliance scoping Specialist vertical agencies
"Let's debate the strategy" buyers Traditional consulting firms
Markets outside France (V1 only) Wait for expansion phases

Promises we make vs. promises we don't make

We promise We don't promise
Outputs: site live, GBP optimized, posts published, SEO work executed, strategy doc shipped Outcomes: revenue lift, conversion rates, ranking positions, follower growth, customer acquisition for the customer's business
Same-day delivery on Website + GBP setup Marketing / SEO measurable impact before 3–6 months
Always-on AI support; business-hours founder escalation Instant human response 24/7
Transparent pricing; no contract lock-in; clear exit terms Free trials beyond Founders pricing (TBD)
Honest scope boundaries (this document) Anything outside this document

How this doc is used

  • Marketing copy — every claim must trace to a "We deliver" cell or "We promise" cell. If it doesn't, it's out.
  • Sales pages / landing page — feature lists pull from delivery columns; "Not for you" sections pull from customer-fit table.
  • Onboarding agent — answers "do you do X?" questions by checking this table first.
  • FAQ — generated directly from the "We don't deliver" + "Categorically outside" columns.
  • Pricing page disclaimers — pull from "We don't promise" column.

When V2 ships, this doc is the first thing updated.

Customer Profiles (ICPs)

Four Ideal Customer Profiles, vertical exclusion principle, launch-focus order.

Source: ~/claude-sync/business/products/ai-agency/customer-profiles.md

AI Digital Agency — Customer Profiles (ICPs)

Locked Session 3, 2026-05-08. Defines who we sell to, who we don't, and where launch focus lives. Marketing channels, ad targeting, content seed verticals, and onboarding-agent personality all derive from this doc.

Vertical exclusion principle (the filter)

Verticals where brand design is the primary buying criterion are NOT our ICP. If the customer's customer judges them on visual taste before anything else, our template-from-gallery + same-day delivery undersells them and breaks the wedge.

Excluded: Hospitality (restaurants, cafés, B&Bs, boutique hotels), fine dining, boutique retail, wedding industry, fashion brands, personal-brand creatives, luxury spa/wellness.

Allowed: Anywhere proof + function beats design — trades, standard wellness, services, auto, pet, education, healthcare (with compliance caveats).

Profile framework

Every ICP defined with these fields:

  1. Vertical examples
  2. Size (employees / scale)
  3. Online maturity (what they have today)
  4. Pain (why they'd buy)
  5. Current alternative
  6. Why we fit (which services map to their pain)
  7. Risk / caveat

ICP 1 — Wellness & Beauty Operator ⭐ BULLSEYE

Field Detail
Verticals Hair salons, barbers, nail salons, beauty institutes, standard spas, massage therapists, tattoo studios
Size 1–10 employees, owner runs the floor
Online maturity Has IG (badly), GBP semi-optimized, often no website or just a Linktree
Pain Customers expect IG presence; reviews drive bookings; owner has zero time after running the chair
Current alternative Doing IG themselves at 11pm, or paying a freelancer €200–400/mo for content only
Why we fit Bundle is exactly their shopping list — Site + GBP + Reviews + Content all matter equally. €79/mo replaces both the freelancer AND the time-debt.
Caveat Visual quality bar on IG output has to clear. Highest LTV potential of all ICPs.

Why this is the bullseye:

  • Highest bundle fit — they buy all four services, not just two
  • Highest LTV potential — visual + reviews + content all matter for their lifecycle
  • Cleanest before/after marketing — visual results sell themselves on IG (our own marketing channel)
  • Decision velocity — owner-buyer, fast yes/no, no procurement
  • Density — France has thousands of independent salons / spas / studios
  • Social-proof flywheel — these customers ARE on IG; they'll post about us if we deliver

ICP 2 — Local Service Trade

Field Detail
Verticals Plumbers, electricians, handymen, locksmiths, mobile mechanics, cleaning services, gardeners, pool maintenance
Size 1–5 employees, owner-operator
Online maturity Unloved GBP, maybe a 5-year-old WordPress someone's nephew built. Phone-and-word-of-mouth.
Pain "Phone rings less than it used to." Younger competitors with reviews and clean GBPs taking jobs.
Current alternative Nothing, or Pages Jaunes listing, or one-time freelancer site (~€800)
Why we fit GBP + Reviews engine maps directly to their pain — reviews drive trust for trades. Same-day Website removes objection. €79/mo bundle = one job's revenue.
Caveat Low IG/FB engagement ceiling — content tier underperforms. Often unbundles to GBP+Reviews+Site = €58/mo.

ICP 3 — Auto Service & Repair

Field Detail
Verticals Independent garages, mechanics, body shops, tire centers, quick-lube, mobile mechanics
Size 2–10 employees, owner-mechanic
Online maturity Old GBP they ignore, maybe a 2010-era site, no IG, no reviews strategy
Pain Reviews drive trust hard (people don't gamble on mechanics); younger competitors with clean GBP + strong reviews steal repeat business
Current alternative Nothing, or a one-off €600 site, or a Pages Jaunes listing
Why we fit Bundle hits hard — Site (functional aesthetic acceptable), GBP + Reviews engine (linchpin), Content (before/after car shots, maintenance tips, customer stories on FB), SEO Basic. Job revenue €100–1,000 each → €79/mo pays back in one customer.
Caveat Lower IG engagement than wellness; FB-heavy market. Older owners → onboarding agent UX must be very simple.

ICP 4 — Independent Professional / Solo Pro

Field Detail
Verticals Coaches, consultants, photographers (non-wedding), personal trainers, fitness instructors, freelance creatives, tutors
Size 1–2, owner IS the brand
Online maturity Half-built Wix abandoned 8 months ago, or LinkedIn-only, or Linktree
Pain Needs a credible site to be taken seriously; can't afford agency; doesn't want to fight Wix
Current alternative Wix/Squarespace DIY (abandoned), or freelancer site that aged badly
Why we fit Website + Strategy + sometimes SEO Basic. À la carte buyers more than bundlers.
Caveat Lower budget ceiling (~€30–60/mo), doesn't always need GBP, churn risk higher. Volume play, not LTV play.

Adjacent fits (worth catching, not chasing)

Vertical Why it works Why not a primary ICP
Pet services (groomers, daycare, training, walking) Reviews-critical, IG-friendly content, low design pressure Smaller LTV per customer
Local education / training studios (driving, music, dance, martial arts, language) Reviews-driven, content-friendly, moderate design demand Niche-by-niche density splits the marketing message
Healthcare (kinés, dentists, vets) Trust + reviews-heavy, reliable budget Compliance caveats limit content claims
Modernizing small family business (auto secondary, optician, florist, bakery) Bundle fit, owner buys Tech-literacy bar makes onboarding harder; uneven IG potential
Childcare / daycare Reviews + local + content fits Compliance + sensitivity around content

Excluded — explicitly not our ICP

  • Hospitality (restaurants, cafés, brasseries, B&Bs, gîtes, boutique hotels) — design-led
  • Fine dining — design-led
  • Boutique retail — design-led
  • Wedding industry (planners, venues, wedding photographers) — design-led
  • Fashion brands & personal-brand creatives — design-led
  • Luxury spa / wellness (high-end tier) — design-led
  • E-commerce-first businesses — out of scope at V1 (no checkout)
  • B2B SaaS / tech — LinkedIn-first market, V1 is IG + FB
  • Multi-location chains — V1 is single business / single GBP / single site
  • Enterprise — no SSO, no procurement, no compliance audits
  • Heavily regulated verticals without compliance scoping (legal, finance) — content compliance walls

Launch focus order

# ICP Role
1 Wellness & Beauty ⭐ Bullseye — full-bundle fit, highest LTV, social-proof flywheel
2 Local Service Trade High-value pain match (reviews + GBP), often unbundles
3 Auto Service & Repair Bundle fit, fast payback, less competitive marketing space
4 Independent Professional / Solo Pro Volume play, lower LTV, à la carte buyer

ICP 1 (Wellness) gets concentrated launch attention. ICP 2 (Trades) and ICP 3 (Auto) compete for second-wave focus — Trades has wider density, Auto has cleaner per-customer economics.

How this doc is used

  • Marketing channels — channel strategy is built around ICP 1 first (where do salon owners hang out?), then ICP 2 and 3
  • Ad targeting — vertical filters, location filters, page-likes filters all derive from these profiles
  • Content seed map — vertical-specific articles, posts, examples follow ICP order
  • Brand voice — calibrated to speak to ICP 1 first without alienating 2/3/4
  • Onboarding agent script — vertical-specific intake questions and recommendations follow these profiles
  • Pricing message framing — "€79/mo replaces your freelancer + time" is ICP 1 framing; "€79/mo = one job's revenue" is ICP 2/3 framing
  • Disqualification messaging — "Not for you if…" sections on landing pages map to the Excluded list

Customer Personas

Sophie (deep), Marc + Karim (sketches). TO-VALIDATE blocks. Validation plan.

Source: ~/claude-sync/business/products/ai-agency/customer-personas.md

AI Digital Agency — Customer Personas (Hypotheses)

Drafted Session 3, 2026-05-08. These are hypotheses, not validated personas. We have zero customer interviews at this point. Each persona ships with a TO VALIDATE section listing what we need to learn from 5–10 real prospects post-launch to confirm or correct.

This doc feeds: ad copy tone, visual direction, channel choice, onboarding agent voice, content topics, objection handling, and pricing message framing.

PERSONA 1 — Sophie Moreau (ICP 1, Wellness & Beauty) ⭐ DEEP

The bullseye — full profile because launch creative + onboarding agent voice + bullseye ad sets all calibrate to her.

Snapshot

  • Age: 36
  • Location: Tours (or any French city 50–250k pop — Reims, Caen, Angers, Annecy)
  • Family: Married, two kids (5 and 9)
  • Business: "Studio Sophie" — owns and runs a hair salon, 2 employees (one stylist + one apprentice), 7 years in business
  • Revenue: ~€110k/year (typical small French salon)
  • Tech literacy: Medium. Uses Planity for bookings, Instagram personally, posts on the salon IG inconsistently, GBP exists but barely touched. Hates dashboards.

A typical day

  • 9:00 — opens, sets up, first client at 9:30
  • 9:30–13:30 — back-to-back appointments
  • 13:30–14:30 — lunch (eats a salad while doing admin)
  • 14:30–19:00 — more clients
  • 19:00 — closes, sweeps up, does the till
  • 20:00 — home, kids, dinner
  • 21:30–22:30 — couch, scrolls IG/TikTok, checks the salon IG, cringes at her last post, swipes Instagram-saves of competitor salons doing it better

The marketing window is 21:30–22:30 Sunday–Thursday. This is when she sees ads, saves them, thinks about them, decides whether to message a sales person or not. It's almost never during business hours.

Goals

  • Fill the chair every day, especially Mondays and Wednesdays (the soft days)
  • Grow the IG to 5k+ followers (vanity but also discovery + booking conversion)
  • Keep up with the new salons opening in town
  • Eventually open a second location, maybe with her cousin — but that's a 2-3 year dream

Pains & frustrations

  1. Bookings are getting less predictable — used to be 80% full, now 65%. Younger clients are going to newer salons with better IG.
  2. She has no time for content. She wants to post but at 21:30 she's exhausted, and what she does post feels lazy and gets 30 likes.
  3. Reviews trickle in but slowly — and the one negative review from last year still haunts her.
  4. Her website is from 2019, made by a freelancer who ghosted her. It's ugly. She knows. She's been "going to fix it" for 18 months.
  5. GBP is half-set-up. She doesn't know what "categories" or "services" or "Q&A" really do.
  6. She has tried Wix once and quit after 3 hours. "Je suis nulle en informatique" (she isn't, she's just busy).

Objections (why she says no)

  1. "Sounds too good to be true." ⭐ The dominant one. €79/mo for site + GBP + content + reviews? Either it's a scam or the quality is garbage.
  2. "L'IA va faire des trucs génériques." Will my IG end up looking like every other AI-generated salon page?
  3. "Je veux garder le contrôle de ma marque." Can I edit? Can I say no to a post?
  4. "Et si je veux partir plus tard?" What if I want to leave — do I lose everything?
  5. "Pourquoi pas Caroline qui me fait du freelance ?" I have a freelancer for €250/mo who does my IG. (Even if she barely posts.)
  6. "Ça va prendre du temps que j'ai pas." The setup will eat my evenings.

Triggers (what makes her actually act)

  1. Sunday night blues + empty Monday calendar → moment of "I have to do something"
  2. Competitor's IG post lands on her feed → "Why does theirs look this good?"
  3. A friend mentions an agency or tool that worked → social proof unlocks her
  4. Founder pricing FOMO ("first 50 at €59/mo") — she's the type who responds to scarcity that feels fair
  5. Same-day delivery promise — immediate gratification matches her impulse-decision style: "ce soir j'ai un site"
  6. A negative review or losing a regular client → urgency spike

Language she uses

  • "le salon," "mes clientes," "un shooting," "un post," "Planity," "réservation"
  • Informal, occasional English ("c'est trop nice," "le before/after")
  • Talks about her business in personal terms ("mon salon," not "the company")
  • Doesn't use words like "ROI," "funnel," "conversion" — would tune out an ad that uses them

Channels (where to reach her)

  • Instagram (heavy daily) — Reels, Stories, Saved posts. Primary channel.
  • TikTok (consumer, not creator) — watches salon content tutorials. Discovery channel.
  • Pinterest — occasional, more for personal style than business
  • Google Search — when she's actively shopping ("logiciel pour salon," "agence digitale prix")
  • WhatsApp — peer recommendations from other salon owners, family
  • Facebook — local commerce groups in her city ("Commerçants Tours")
  • NOT LinkedIn. NOT Twitter/X. NOT email newsletters from strangers.

Decision-making style

  • Emotion + social proof + impulse, in that order.
  • Sees an ad → saves it → 2-3 days of revisiting → asks one trusted friend → decides on impulse
  • Won't watch a 20-min webinar. Will watch a 30-sec testimonial reel.
  • Trust is built by: real visible salons we've helped (with names + photos), founder face on camera, "no contract" + "satisfaction guarantee" + "cancel anytime"

Money attitude

  • €79/mo = "le prix d'une coupe-couleur." She can rationalize this against one client per month.
  • Annual prepay (2 months free) is attractive but she'll hesitate — most salons cash-flow month-to-month.
  • Above €100/mo and she'll second-guess. Below €50/mo and she'll think it's cheap = bad quality.
  • The €79 sweet spot is psychologically right for her.

What she reads / watches / who she trusts

  • IG: top French hair brands & influencers (e.g., L'Oréal Pro, Kérastase, well-known French stylists)
  • TikTok: salon tutorials, transformation reels
  • Magazines: occasional ELLE, Marie Claire — but mostly IG-mediated
  • Trade press: occasional (Coiffure de Paris, Biblond)
  • Her trade community: other salon owners in her city, Planity content, friends
  • Will trust: another salon owner who's used us (testimonial), local-named businesses, transparent founders
  • Will distrust: glossy stock photography, generic "agency" branding, "AI" as a hype word

TO VALIDATE (post-launch interviews)

  • Is the marketing window really 21:30–22:30, or is it Sunday morning / Saturday afternoon?
  • Is "trop beau pour être vrai" the dominant objection, or is "I want control" actually first?
  • Does the freelancer-replacement angle land, or do they not even have a freelancer?
  • Is €79/mo the right anchor, or does framing it as "le prix d'une coupe-couleur" actually feel cheap to them?
  • What language do they use for our services in their own words? (Their words > our words for ad copy.)
  • How much do testimonials from other salons matter vs. testimonials from any small business?

PERSONA 2 — Marc Dubois (ICP 2, Local Service Trade) — SKETCH

Snapshot

  • Age: 47
  • Location: Outer-ring suburb of Lyon (or any French metro)
  • Business: Independent plumber, SARL, 12 years in business, 1 employee (apprentice)
  • Revenue: ~€85k/year
  • Tech literacy: Low-medium. WhatsApp heavy, Pages Jaunes still relevant in his head, smartphone-first.

Top 3 pains

  1. The phone rings less. Younger people Google + read reviews now; he has 14 reviews and a 4.1 rating from 2019.
  2. His website (built by a cousin in 2018) is broken on mobile, and he knows it but doesn't know how to fix it.
  3. A new local plumber with a slick GBP and 80 reviews is taking jobs in his area.

Top 3 objections

  1. "Internet, c'est pas pour moi." Tech avoidance — needs the onboarding agent to feel like a phone call, not a form.
  2. "€79 par mois, c'est cher pour moi." Even if one job pays it back, monthly recurring is psychologically harder for trades than one-shot.
  3. "Je connais déjà mes clients, j'ai pas besoin d'Internet." Doesn't yet feel the loss is digital — blames "the economy."

Buying trigger

  • A 2-week stretch of slow phone activity. Or his apprentice mentioning "you should do TikTok like Pierrot" (and him hating that idea but realizing it's true).

Channels (to reach him)

  • Facebook (heavy, personal + local commerce groups)
  • Google Search ("comment avoir plus de clients plombier")
  • WhatsApp peer chat
  • YouTube for trade content
  • NOT Instagram (he scrolls but doesn't post)
  • NOT TikTok (visceral aversion)

Onboarding voice

Plain, direct, no jargon. Explain things like he's a friend. Avoid "AI" as a hype word — say "automatisé" or just describe what happens.

TO VALIDATE

  • Does monthly subscription actually scare him, or is it just the price?
  • Is FB the right ad channel, or should we be in trade-specific groups (Facebook groups for "Artisans Plombiers France" etc.)?
  • Does "1 client = €79" framing land, or does he need a different anchor?
  • Will he want to talk to a human before subscribing, or will an onboarding agent that "feels like a phone call" be enough?

PERSONA 3 — Karim Benali (ICP 3, Auto Service & Repair) — SKETCH

Snapshot

  • Age: 52
  • Location: Suburb of Paris / Lyon / Marseille
  • Business: Independent garage, 22 years in business, 3 mechanics employed, family-run (wife does books)
  • Revenue: ~€280k/year
  • Tech literacy: Low. Uses smartphone for WhatsApp + calls + GBP notifications. His nephew set up the GBP.

Top 3 pains

  1. Reviews matter more every year, and he has 28 of them. The new garage down the road has 200+.
  2. People shop around online before bringing the car in — he's losing first-time customers to garages with better online presence.
  3. He doesn't trust marketing agencies. Got burned once for €400 with no result.

Top 3 objections

  1. "J'ai déjà été arnaqué par une agence." ⭐ The dominant one. Trust trauma.
  2. "Internet, c'est pour les jeunes." Genuinely thinks digital marketing isn't for his market.
  3. "Mes clients, ils m'appellent, ils viennent pas par Google." Underestimates the share of first-time customers who do.

Buying trigger

  • His son or nephew shows him a competitor's GBP with 200 reviews and beautiful photos. Family pressure beats ad pressure for this persona.
  • Or: a customer mentioning "j'ai failli aller au garage X mais leurs avis étaient pas terribles" — direct feedback that reviews lost or won the customer.

Channels (to reach him)

  • Facebook (heavy)
  • WhatsApp (very heavy — group chats with family, with peers, with regulars)
  • YouTube (mechanic content)
  • NOT IG, NOT TikTok, NOT LinkedIn
  • Word-of-mouth via family members is the strongest channel. If his nephew or his son recommends us, he'll listen.

Onboarding voice

Respectful, plain, direct, family-warm. Explain things like a son explaining tech to a father. No condescension. Phrase pricing as "le prix d'un changement d'huile" or "moins qu'une révision" — bring it into his world.

TO VALIDATE

  • Is family-recommendation really the dominant trigger, or is it competitor pain that breaks the camel's back?
  • Does the "agency trauma" objection require an explicit money-back guarantee in copy, or is "no contract, cancel anytime" enough?
  • Will he subscribe online via the onboarding agent at all, or does he need a phone call to close? (This is the biggest question — if Persona 3 needs human-closed sales, that breaks our scaling model for this ICP.)

ICP 4 (Independent Professional / Solo Pro) — INTENTIONALLY NOT PERSONA'D

Caught as inbound during launch. Persona work happens when this segment becomes a focus wave. Until then, the landing page just needs to not exclude them and the onboarding agent needs to handle their (lighter) configuration without bespoke vertical branching.

How this doc is used

  • Ad creative — every ad we run names which persona it targets and which trigger/objection it leans into
  • Onboarding agent voice — vertical branch in the agent's tone (Sophie's branch is lighter/warmer; Karim's branch is plainer/more respectful)
  • Channel budget split — IG/TikTok-heavy for Sophie, FB + Google + family-network for Marc and Karim
  • Content seed map — articles and posts targeted at each persona's pains and search queries
  • Pricing page copy — anchors map to each persona's "money attitude" line
  • FAQ entries — generated directly from the objections lists
  • Disqualification copy — pulls from the customer-fit table in customer-profiles.md, but tone calibrated by persona

Cross-doc dependencies

  • customer-profiles.md — defines WHICH businesses; this doc defines WHO inside the business buys
  • scope-boundaries.md — defines WHAT we can deliver; objection responses must trace back to it
  • business-case.md — wedge claims must land in the language each persona uses
  • lean-plan.md — pricing math; persona "money attitude" calibrates how we phrase it

Validation plan (post-launch)

  • Target: 5–10 customer interviews per persona within 60 days of launch
  • Format: 30-min video calls, semi-structured around each persona's TO VALIDATE list
  • Output: persona update commit with corrected fields and removed assumptions
  • Trigger: when a customer subscribes, automatic 7-day-later email asking for a 30-min interview slot, €30 thank-you credit on their next month

Lean Business Plan

Per-service cost-to-deliver, pricing structure, marketing operations, dependency map.

Source: ~/claude-sync/business/products/ai-agency/lean-plan.md

AI Digital Agency — Lean Business Plan

Session 2 of the co-founder workstream. Builds on business-case.md (LOCKED Session 1, amended Session 2 kickoff).

Working assumptions (this session)

  • Standalone-business framing (locked 2026-05-08): This business is modeled as a standalone venture with its own readiness requirements, not as a fix for any personal financial situation. Launch happens when the business is ready (product, brand, marketing, capital), not when external pressure says it should. If the business eventually solves personal financial problems, fine — but that's never a design constraint and never a launch trigger. Breakeven, projections, and milestones are framed against the business's own cost structure (cost-to-deliver + marketing budget + tooling), not against any personal monthly burn.
  • Founder hourly rate (planning placeholder): €20/hour ≈ 65 DT/hour. This is an internal cost number — what we charge our own time at when computing per-customer cost-to-deliver. Not a billable rate. Adjust anytime.
  • API runtime: Claude API from day 1 (paid). Per-customer token cost is a real line item.
  • Founder split-attention: holds. Lean plan models part-time founder availability, not full-time.
  • Currency for plan numbers: € (because pricing benchmarks come in € and majority of pipeline is FR/EU). DT equivalents on demand.
  • Customer lifetime for amortization: 12 months as a planning baseline. Adjustable when we get real churn data.

Cost-to-deliver per service

For each service we capture:

  1. One-time setup cost (founder hours + AI tokens + tools) → amortized monthly over assumed customer lifetime
  2. Ongoing monthly cost (founder hours + AI tokens + tools / pass-through)
  3. Total cost-to-deliver per service per customer per month

IMPORTANT framing: these numbers are per service, not per customer. À la carte customers only carry the cost of the services they bought.

Per-customer total cost-to-deliver = (sum of services bought) + (24/7 support inclusion overhead, counted once per customer regardless of services).

Plus 24/7 customer support as a separate cost line (table-stakes inclusion, attached to every customer once, not multiplied by services bought).

Service 1 — Website (same-day delivery, brochure-only)

Workflow (per Boss, 2026-05-07):

  1. AI onboarding agent interview — collects all required info: text content, images, contact info, legal info, services list, pricing, hours, etc.
  2. Mockup gallery — customer browses our mockup gallery and picks a style. If customer provides a logo, the system shows recommended mockups that pair well with the logo.
  3. Website building framework — mockups have predefined pages, text blocks, and SEO structure (headings, meta tags). AI populates content into the mockup slots from the customer's intake answers. Output is basic-SEO-ready (proper H1/H2 hierarchy, meta descriptions, alt text).
  4. Customer review — customer reviews the populated site. They can:
    • Request AI to regenerate any block
    • Manually edit (with a cautious in-product warning about SEO impact)
  5. Customer approves final version
  6. Domain + hosting + SSL setup
    • Either customer buys domain through us (we resell), OR
    • Customer brings own domain → we send DNS instructions
  7. Push live

No founder intervention in the normal flow. Founder only steps in for escalations (DNS that won't resolve, customer hits a limit AI can't handle, etc.).

Possible add-on: Logo design (paid).

One-time setup we owe before launch (not per-customer cost): convert the existing template ZIPs on this VPS into a previewable mockup gallery.

Decisions locked Session 2:

  • Q1 (ongoing edits) → option (a): AI executes capped edits, customer approves, founder = 0 min in normal flow. Escalations only.
  • Q2 (domain when bought through us) → option (b): retail markup. Wholesale ~$10/year, retail €15–20/year, margin ~€5–10/year per customer.
  • Q3 (AI onboarding agent token accounting) → CAC line, separate from any service. ~$3–5 per converted customer (token + OTP) after defenses.

AI onboarding agent — defense stack locked Session 2:

  • Phone OTP-first gate before any AI tokens are spent. Trumps email-first because a real number is much harder to fake than a disposable email.
  • Provider: Twilio, $0.05/SMS planning placeholder (revisable when Q11 launch market locks — Tunisia/France ~€0.04–0.08, SN/CI ~€0.08–0.12).
  • Visitor-facing explanation strip at the OTP gate: explain why we ask (fraud / token-burn protection). Boss insight: SMB owners get fake orders all the time and immediately relate to the reasoning — converts friction into shared context.
  • Cloudflare Turnstile as a pre-OTP filter (kills bots before we pay for SMS).
  • Haiku 4.5 for the chat conversation. Sonnet 4.6 only for the final quote-PDF burst, after intake completes.
  • Per-session, per-IP, and global daily token caps as backstops.
  • Architecture full spec → PRD session. Lean-plan only carries the cost number.

Final Service 1 cost-to-deliver (locked Session 2):

Component One-time / Monthly Cost
AI tokens — content population (Haiku for slot-fill + Sonnet for hero/about copy, FR/AR/EN, 1–2 regeneration cycles) One-time ~€3
Founder escalation avg (DNS issues, edge cases) One-time ~€1.70 (5 min @ €20/hr)
Mockup gallery browse, hosting setup, SSL, push live One-time €0 (automated)
Domain (when bought through us) One-time Net €0 — wholesale ~$10 covered by retail markup
Setup total per customer ~€5
Setup amortized over 12-mo lifetime Monthly ~€0.40
AI tokens — capped edits (~2/mo avg, AI-executed) Monthly ~€0.40
Founder escalation avg Monthly ~€1.70 (5 min @ €20/hr)
Hosting share on existing VPS Monthly €0.50–1
Domain renewal (when applicable) Monthly Net €0 — markup covers wholesale
Ongoing total per customer Monthly ~€2.60–3.10
TOTAL cost-to-deliver Website per customer per month ~€3

Range: €2.50–3.50/month all-in (setup amortized + ongoing).

Why this is so much lower than the first draft (€13): my first model had founder QA time baked into the build flow. Boss corrected: there is no founder QA in the normal flow. The workflow is fully customer-facing AI; founder only handles escalations.

Service 2 — GBP (Google Business Profile)

Integration decision (locked Session 2): Build our own MCP wrapping Google's GBP API directly. Rejected Zapier (only 3 actions exposed, expensive at scale) and Pipedream (same middleman trade-off). Owning the integration is part of the wedge — Zapier/Pipedream costs scale linearly with customer count, ours is one-time.

Build cost: €0 — Boss's Claude Max plan covers the build (Alfred writes it). Boss's involvement: ~30 minutes to kick off + review + deploy.

Prerequisite: Google production API access approval. Multi-week review. Notion task created to start the application now: "Request Google Business Profile API access for production use" (Tech, High priority).

GBP API surface available (rich — full CRUD on locations, reviews, posts with recurring schedules, direct photo upload, Q&A, verifications, performance/insights). Anything Boss wants the AI to do on GBP is technically possible.

Workflow (per Boss, 2026-05-07):

Path A — customer has existing GBP:

  1. AI walks customer through standard Google OAuth consent flow — they grant our app access via Google's screen, never share their password.
  2. Once authorized, AI audits the listing, fills missing fields, normalizes NAP, suggests categories, drafts an initial post.
  3. Customer reviews initial state in our dashboard. Live.

Path B — customer doesn't have GBP yet:

  1. AI guides customer through written + video guide to create the GBP and complete Google's verification (postcard / phone / video — whichever Google offers).
  2. Customer self-serves through the guides; if stuck, opens a support request. (No founder shadowing on a call by default.)
  3. Once verified, customer hands over via OAuth — same as Path A from here.

Ongoing cadence (Q-GBP-2):

  • Both auto and manual modes offered. Customer picks per-feature (posts, review replies, listing updates).
  • Manual mode: AI drafts → owner approves in dashboard → AI publishes.
  • Auto mode: AI drafts → AI publishes directly. Sold as a premium tier — pricing lever.

Photos (Q-GBP-3): Customer photo bank uploaded at intake. AI tells them what to shoot (format, lighting, what subjects, recommended quantity per category). Monthly: AI rotates from the bank; nudges customer when bank runs low.

Pre-launch one-time builds owed (not per-customer cost):

  • GBP MCP wrapping the Google API (Alfred, ~30 min Boss time)
  • Written verification guide (Path B)
  • Video verification guide (Path B)
  • Photo-shooting recommendation guide (intake)

Service 2 cost-to-deliver:

Component Path A (existing GBP) Path B (new GBP)
OAuth flow / verification handholding (founder time) 5 min escalation avg = €1.70 10–15 min avg (verification support) = €4
AI tokens — initial audit + fixes + first post (Haiku/Sonnet mix) ~$0.50 ~$0.70 (extra support tokens)
Setup total ~€2.20 ~€4.70
50/50 blended setup ~€3.50 per customer
Setup amortized over 12 mo ~€0.30/mo

Ongoing monthly (per customer, mode-agnostic — token cost is the same whether AI publishes directly or after approval click):

Line Cost / customer / month
AI tokens — posts (~4/mo) ~€0.10
AI tokens — review replies (~5/mo avg) ~€0.05
AI tokens — monthly listing health check + drift fixes ~€0.15
AI tokens — photo handling (rotate, alt text, monthly nudge) ~€0.10
Founder escalation avg ~€1.70 (5 min @ €20/hr)
Ongoing total ~€2.10/mo

Cost-to-deliver Service 2 — GBP: ~€2.50/month per customer

(setup amortized €0.30 + ongoing €2.10)

Range: €2.30 (mostly Path A — easier flow) to €2.80 (mostly Path B — verification handholding eats more founder minutes).

Pricing implication worth flagging now: "Auto mode" costs us the same as "Manual mode" to run (the tokens are identical — AI generates the post either way). Selling Auto as a premium tier is near-pure-margin revenue. If Auto adds €5–10/mo at retail, on a €2.50 cost base, that's a margin lever we should absolutely use.

Service 3 — Review Engine

Workflow (locked Session 2):

Trigger: manual (owner-initiated) + passive (QR sticker / plexi). No POS/CRM integrations in V1.

Send mechanism — the wedge move: WhatsApp + SMS deep-link via owner's own phone.

  • Our PWA opens the owner's default SMS or WhatsApp app pre-populated with recipient + message template.
  • Owner taps Send once. Message goes out through owner's carrier on owner's plan with owner's number as sender.
  • Cost to us per send: €0. Cost to owner: covered by their existing plan (effectively €0 for postpaid plans in our markets).
  • Bonus: recipient sees a number they recognize → higher response rate than a generic gateway number.
  • The review link inside the message is OURS (https://review.{customer-domain}/r/{token}) — funnel intelligence stays intact even though the message itself is off our infrastructure.

Channels offered V1:

  • WhatsApp deep-link — primary in our markets (TN/FR/MA/SN/CI all WhatsApp-heavy)
  • SMS deep-link — fallback for non-WhatsApp recipients
  • Email — free via our SMTP
  • QR (sticker or plexi) — passive, paid add-on with margin

Plexi / sticker as paid add-on (locked):

  • Both options offered. Customer picks format and quantity at intake or as upsell.
  • Pass-through with retail markup (same model as domain). Wholesale ~1-3 DT (sticker) / 10-30 DT (plexi); retail ~30-50 DT (sticker pack) / 80-150 DT (plexi piece). Margin ~20-40 DT per add-on per customer.

Premium upgrade (later, not V1): "Bulk send via Twilio gateway" — for high-volume customers (busy clinics, big salons, high-traffic trades) who don't want to tap Send manually. Cost ~€0.05/SMS, retail ~€0.10/SMS = 50% margin. Volume play.

Review monitoring: uses same GBP OAuth scope as Service 2. If a customer buys Review Engine but NOT GBP service, they still grant OAuth the same way — just for review-read access. No extra integration cost.

Defaults applied (flag for Boss confirmation):

  • Reply mode: same pattern as GBP — both auto and manual offered, auto sold as premium tier.
  • Negative reviews (≤2 stars): AI drafts a response but always waits for owner approval regardless of mode. Auto-replying to a 1-star is a tone landmine; never let it auto-fire.
  • Monthly velocity report: dashboard view + monthly PDF email. AI-generated, low cost.

Service 3 cost-to-deliver:

Component Cost
Setup
Founder escalation avg (PWA install help, OAuth issues) ~€1.70 (5 min @ €20/hr)
AI tokens — onboarding chat, template generation per language, first welcome ~€0.30
Setup total ~€2 per customer
Amortized over 12 mo ~€0.17/mo
Ongoing monthly
AI tokens — message drafting (~30 sends/mo avg, Haiku) ~€0.30
AI tokens — review monitoring + classification (daily batch) ~€0.15
AI tokens — reply drafting (~5 reviews/mo avg) ~€0.10
AI tokens — negative review alerts ~€0.05
AI tokens — monthly velocity report (Haiku + Sonnet polish) ~€0.20
Per-message cost (deep-link mode) €0
Hosting share (review tracking links on existing VPS) ~€0.20
Founder escalation avg ~€1.70 (5 min)
Ongoing total ~€2.70/mo

Cost-to-deliver Service 3 — Review Engine: ~€2.85/month per customer

(setup amortized €0.17 + ongoing €2.70)

Range: €2.50 (low review volume) to €3.20 (high review volume — clinics, restaurants).

Pricing levers exposed by this service:

  1. Plexi/sticker margin — 20-40 DT per add-on per customer. Real revenue line, not cost.
  2. Auto-reply mode — premium tier upsell, same near-pure-margin lever as GBP.
  3. Bulk gateway upgrade — for high-volume customers later. ~50% margin per SMS, volume play.

Service 4 — Content (social media)

Workflow & architecture (locked Session 2):

  • V1 platforms: Facebook + Instagram (default for all customers) + LinkedIn (gated to B2B verticals only) — AI onboarding agent asks "do you serve businesses or individuals?" and routes content accordingly. Don't waste LinkedIn token cost on a salon owner.
  • TikTok deferred to V2 — video-first paradigm, requires AI video generation (paid APIs) that pairs with the V2 image-gen wave.
  • Posting infrastructure: Postiz (self-hosted, AGPL-3.0). Decided over Composio after research:
    • Composio's open-source repo is just SDKs — the runtime is their cloud SaaS. "Self-hosted" Composio = enterprise VPC, sales-only.
    • Composio cloud pricing: $29/mo at 200K tool-calls (we'd hit this around 200 customers), $229/mo at 2M tool-calls — recurring forever.
    • Postiz: fully self-hostable, no phone-home, all platforms in OSS version. €0 marginal cost forever, ~1–2 GB RAM additional on VPS.
    • Same "own the integration" play as GBP MCP. Architecturally coherent.
    • Postiz exposes an API we can wrap in our own MCP if AI agents need direct access — €0 build under Boss's Max plan.
  • Visuals V1: customer photo bank only (same bank as GBP), with programmatic text-overlay templates. €0 image cost.
  • AI image generation: V2 — pairs with TikTok wave.
  • Auto/manual mode: same pattern as GBP/Reviews. Both modes per-feature (post publishing, comment replies, DM replies). Auto = premium tier. Hostile/negative comments always wait for owner approval regardless of mode.

Pricing tiers built into the service (cadence-driven):

  • Low: 4 posts/mo
  • Mid: 16 posts/mo (reference tier for cost math)
  • High: 30+ posts/mo

Service 4 cost-to-deliver — mid tier (16 posts/mo, 2.5 platforms blended):

Component Cost
Setup
Founder escalation avg (Postiz OAuth, photo bank import) ~€1.70 (5 min @ €20/hr)
AI tokens — 3-month content calendar (Sonnet) + first 4 posts (Haiku) ~€0.55
Setup total ~€2.25 per customer
Amortized over 12 mo ~€0.19/mo
Ongoing monthly (mid tier, 16 posts/mo)
AI tokens — base post generation (Haiku) ~€0.20
AI tokens — platform adaptations (FB/IG/LinkedIn voice tuning, Haiku) ~€0.18
AI tokens — hashtags, captions, alt text ~€0.15
AI tokens — comment/DM auto-reply (~30 interactions/mo, Haiku) ~€0.24
AI tokens — monthly content calendar refresh (Sonnet) ~€0.45
AI tokens — monthly performance report ~€0.20
Per-publish cost (Postiz self-hosted) €0
Hosting share (Postgres + Redis + Temporal + Postiz) ~€0.30
Founder escalation avg ~€1.70 (5 min)
Ongoing total — mid tier ~€3.40/mo

Cost-to-deliver Service 4 — Content (mid tier): ~€3.60/month per customer

Tier breakdown:

Tier Posts/mo Cost-to-deliver
Low 4 ~€2.60/mo
Mid 16 ~€3.60/mo
High 30+ ~€5/mo

Pricing levers exposed:

  1. Tier-based cadence is the natural pricing axis (more posts = more value, more token cost).
  2. Auto-publish + auto-reply premium tier — same near-pure-margin upgrade as GBP/Reviews.
  3. B2B LinkedIn gating is a cost-control lever, not a pricing lever (cuts unnecessary token spend on non-B2B customers).

Pre-launch builds added from Service 4:

  • Deploy Postiz on VPS (~30 min Boss time)
  • Wrap Postiz API in our own MCP for AI-agent calls (€0 build)
  • Vertical-tuned content prompts (one-time prompt engineering across ~10 verticals)
  • Programmatic image-overlay templates (text-on-photo styles, brand-color application)
  • Comment/DM intake routing (sales inquiry → owner; spam → block; troll → ignore)

Service 5 — SEO (two tiers: Basic + Advanced)

Note: GEO (AI visibility / brand citation in ChatGPT/Claude/Perplexity/AI Overviews) separated out as its own service / add-on — to be sized after the core 7. SEO here = traditional Google/local search optimization.

Locked structure:

SEO Basic:

  • 2 articles/mo, AI writes → owner approves → publish
  • Free tracking path: GSC + GA4 (existing MCPs on this VPS) + Google Trends + AI competitive scans
  • Skip citations
  • Schema markup included (table-stakes for the site we built in Service 1)

SEO Advanced:

  • 4 articles/mo, AI writes → founder reviews + edits → customer approves → publish
  • Seobility tool ($59/mo, 3 customer sites per account → ~€18.50/customer/mo at full utilization)
  • Citations available on-demand (separate pricing, not in base recurring)
  • Deeper audit + richer monthly performance report

Feedback flagged by Alfred (Session 2), retained in plan:

  • Founder-review at scale is the hidden bottleneck on Advanced. 80 min/customer/mo × 100 Advanced customers = 133 hr/mo on blog reviews alone. Two paths to address:
    • (b) Price Advanced explicitly to fund the founder time at launch — €120–180/mo retail covers 80 min comfortably.
    • (a) Lighter QA pattern by month 6 — AI runs strict QC pass, founder spot-checks 1 in 4 articles randomly. Cuts founder time ~70% while preserving the "human-reviewed" claim.
    • Recommended: (b) at launch + (a) when volume justifies it.
  • Seobility €18.50/customer share is real recurring cost. Apply pass-through-with-markup pricing (e.g., €30/mo "premium SEO tool access" line baked into Advanced retail price, €11.50 margin per customer).
  • Citations on-demand needs a price list, not "ask us": suggested starting points €5–10/citation single, €100–150 for 30-pack, €150–250 for NAP audit + cleanup.

SEO Basic cost-to-deliver:

Component Cost
Setup
AI tokens — initial site scan, keyword research, baseline recommendations ~€0.80
Founder escalation avg ~€1.70 (5 min)
Setup total ~€2.50 per customer
Amortized over 12 mo ~€0.21/mo
Ongoing monthly
AI tokens — 2 articles (Sonnet outline + Haiku body) ~€0.90
AI tokens — monthly on-page audit + GSC/GA4 read + suggestions ~€0.30
AI tokens — keyword tracking via free tools ~€0.20
AI tokens — monthly performance report ~€0.20
Hosting share ~€0.20
Founder escalation avg ~€1.70 (5 min)
Ongoing total ~€3.50/mo

Cost-to-deliver SEO Basic: ~€3.70/month per customer

SEO Advanced cost-to-deliver:

Component Cost
Setup
AI tokens — deeper competitive scan, keyword universe build, Seobility account setup ~€1
Founder escalation avg ~€1.70 (5 min)
Seobility account provisioning (one-time) €0 (admin task, included in subscription)
Setup total ~€3 per customer
Amortized over 12 mo ~€0.25/mo
Ongoing monthly
AI tokens — 4 articles (Sonnet for outline + polish, Haiku for body) ~€2.20
Founder review time — 4 × 20 min ~€27 (1.33 hr @ €20/hr)
AI tokens — monthly audit + Seobility data integration ~€0.50
AI tokens — monthly performance report (richer with Seobility) ~€0.30
Seobility tool share ~€18.50 (at 3 customers per $59 account)
Hosting share ~€0.30
Founder escalation avg ~€1.70 (5 min)
Ongoing total ~€50.50/mo

Cost-to-deliver SEO Advanced: ~€50.75/month per customer

(Citations on-demand priced separately, not in this number.)

Implication: Advanced is ~14x the cost of Basic. That's load-bearing on the wedge — Advanced needs to retail at ~€120–180/mo to hit healthy margin, vs. ~€30–50 for Basic. The founder-time line is what makes Advanced expensive; the lighter-QA pattern at month 6 cuts that by 70% and changes the unit economics significantly.

Pre-launch builds added from Service 5:

  • Vertical-specific keyword research prompts (~10 verticals)
  • Article generation pipeline (Sonnet+Haiku orchestration, brief → outline → draft → polish)
  • Customer-facing GSC + GA4 dashboard view
  • Seobility account provisioning workflow (3 customers per account)
  • Citation submission templates (for on-demand requests)
  • Schema markup library per vertical (one-time)

Service 6 — Strategy

Workflow (locked Session 2):

  • Format: PDF only (no dashboard).
  • Generation: fully AI. Sonnet generates from customer intake answers + competitor scan. No founder touch in normal flow. Customer receives the PDF; can request a revision if needed.
  • Length: ~10–15 pages, branded template with simple charts (positioning, channel mix, content angles, KPIs, 3-month action plan).
  • Cadence: initial doc at onboarding + automatic refresh every 6 months, included in subscription.
  • On-demand refresh: €10 per refresh. Priced low enough to be a "yes please" purchase, not a deliberation. High attach rate likely.

Service 6 cost-to-deliver:

Component Cost
Setup + 6-mo refresh (counted across 12-mo lifetime)
AI tokens — initial strategy doc (Sonnet for thinking + multilingual: FR/AR/EN) ~€0.85
AI tokens — 6-month refresh ~€0.85
PDF generation (programmatic, branded template) €0
Founder time on initial + refresh €0 (fully AI)
Total over 12-mo customer lifetime ~€1.70
Amortized monthly ~€0.14/mo
Ongoing monthly
AI tokens — small interactions (customer asks questions about their strategy doc) ~€0.10
Founder escalation pool (rare — "where's my doc" type issues) ~€0.50 (1.5 min/mo)
Hosting share ~€0
Ongoing total ~€0.60/mo

Cost-to-deliver Service 6 — Strategy: ~€0.75/month per customer

On-demand refresh economics:

  • Cost per refresh: ~€1 (€0.85 tokens + €0.15 friction)
  • Retail: €10
  • Margin: ~€9 per refresh, near-pure-margin add-on
  • If 30% of customers buy 1 refresh per year → ~€3 incremental revenue/customer/year at virtually no extra cost.

Pricing observation: This is the cheapest service in the catalog by far. Strategy is essentially "free to give" once the prompt and PDF pipeline are built. Useful as a bundle anchor ("includes a personalized digital strategy worth €X") to bulk up the perceived value of the bundle without bleeding margin.

Pre-launch builds added from Service 6:

  • Strategy doc Sonnet prompt (positioning + channel mix + KPIs + action plan, with vertical-specific framings)
  • PDF generation pipeline (markdown → branded PDF, multi-language)
  • 6-month refresh trigger (cron job firing 6 mo after onboarding)
  • On-demand refresh purchase flow

Service 7 — AI Chatbot — DEFERRED to V2

GEO (planned standalone service) — DEFERRED to V2

Decision (locked Session 2): Both AI Chatbot and GEO are out of V1 scope. Group them into the V2 wave with TikTok and AI image generation.

Rationale:

  • AI Chatbot: Only service where token cost is driven by visitors of the customer's site, not by the SMB owner. Different threat model (anonymous internet traffic, friction-vs-leads tension, harder to gate without hurting customer's conversion). Most engineering-heavy pre-launch piece. Not core to the wedge.
  • GEO: Methodologies still maturing in 2026; third-party AI assistant API costs (OpenAI / Gemini for tracking) are real recurring cost lines. Better validated after V1 customer base proves the core wedge.
  • Coherent V2 story: "AI chatbot + GEO + TikTok + AI image gen" packaged together becomes a strong upsell narrative ("Now your business is fully AI-native — visitors get answered 24/7, you're cited in AI assistants, you reach video-first audiences") for existing customers in month 6+.

V1 = 4 core services + 1 add-on + Strategy add-on + 24/7 customer support inclusion:

  1. Website (same-day delivery, brochure-only)
  2. GBP + Reviews merged (Google Business Profile management + Review Engine sold as one product). Combined cost-to-deliver ~€5.35/mo. Reviews-as-GBP-add-on rule made the merge natural; share OAuth, share dashboard, share customer setup. Split back into separate products in V2 when automated review-send (Twilio bulk gateway) becomes the differentiator for a premium tier.
  3. Content (social media: FB + IG + LinkedIn-for-B2B)
  4. SEO Basic only (V1 — Advanced deferred to V2)
  5. Strategy — add-on only (PDF + 6-mo refresh + €10 on-demand). Attaches to Website OR Content (social), never sold standalone.

SEO Advanced → V2 (locked Session 2): Boss explicitly deferred SEO Advanced to V2, gated on hiring. Rationale: at 4 articles/mo with founder review × 100 Advanced customers = 133 hr/mo of founder time on blog QA alone, on top of all other work. Solo founder can't carry it. Re-launches in V2 once cash flow supports a content-QA hire.

  • V2 retail price for SEO Advanced (locked): €99/mo, standalone-only (cannot be bundled).
  • Cost-to-deliver remains ~€50.75/mo with current pattern, drops to ~€31/mo with lighter QA pattern once hire is trained.
  • Margin expectations at V2 launch: ~50% gross (€48 margin on €99 retail) at full pattern; ~70% (€68 margin) once lighter QA lands.

Strategy as add-on only (locked Session 2): Cannot be sold à la carte. Customer must purchase Website OR Content (social media) as the base service; Strategy can then be attached. Rationale: a strategy doc only delivers value when there's something to execute on (a site to drive traffic to, social channels to fill with content). Selling Strategy alone would attract margin-thin customers carrying the full €5.50 support overhead.

SEO Basic as Website add-on only (locked Session 2): SEO Basic requires Website service — we only optimize sites built on our platform. Rationale: our SEO pipeline (schema, on-page templates, technical SEO, sitemap, robots) is wired into our template framework. Auditing and optimizing a customer's existing Wix/WordPress/Webflow site would require parallel engineering for each framework — out of V1 scope. Quality-control move as much as a cost-control move.

Sales-script implications (must be enforced by AI onboarding agent):

  • "Strategy alone" → redirected: "A strategy doc works best when paired with what you'll actually execute on. Want to start with Website + Strategy or Content + Strategy?"
  • "SEO for my existing site" → redirected: "Our SEO is built into our website framework. Same-day rebuild of your site on our platform unlocks the SEO service. Want to see how that works?"

Review Engine as GBP add-on (locked Session 2): Cannot be sold à la carte. Customer must purchase GBP service as the base; Review Engine attaches on top.

Rationale (technical and product):

  • The Review Engine has two halves. The outbound half (sending review requests via WhatsApp/SMS deep-link, email, QR) works without GBP. But the inbound half — review monitoring, AI-drafted reply suggestions, negative-review alerts, velocity tracking — requires GBP OAuth to read incoming reviews and post replies. Without GBP, Reviews degrades to a half-product ("send links and hope") that contradicts the agency-grade quality wedge.
  • The GBP OAuth is the same OAuth Reviews needs. Bundling them means one auth flow, one customer setup, cleaner UX.
  • Eliminates a margin-thin standalone (Reviews-only at €2.85 service / €5.50 support = €8.35 cost vs. Reviews+GBP at €5.35 service / €5.50 support = healthier margin profile).

Sales-script implication: AI onboarding agent redirects "Reviews only" inquiries: "Reviews work best when we manage your Google Business Profile too — same setup, same dashboard. Want to bundle them?"

Service dependency map (V1) — final:

Service Standalone? Depends on
Website ✓ — (anchor 1)
GBP ✓ — (anchor 2)
Content (social) ✓ — (anchor 3)
Review Engine ✗ Requires GBP
SEO Basic ✗ Requires Website
Strategy add-on ✗ Requires Website OR Content
24/7 support auto Attached to any purchase

Architectural pattern — three anchor services, three add-on lanes:

Anchor Dependent add-ons Customer profile
Website SEO Basic, Strategy Web-focused — wants a digital presence + ranking + plan
GBP Review Engine Local-discovery-focused — wants to dominate local search + reviews
Content (social) Strategy Social-focused — already has site, wants managed presence + plan

Customers self-segment into one of three lanes (or combine multiple). The dependency structure pulls every customer toward an anchor service + add-ons inside that lane, naturally building bundle AOV instead of producing margin-thin orphan SKUs.

Real à la carte options at V1: Website / GBP / Content. Three. Everything else is dependent.

Wedge-claim adjustment for next sales-marketing pass: The business case wedge originally listed "site + GBP + review system + social content + SEO/GEO + strategy + AI chatbot" (7 items). V1 wedge reads cleaner as "site + GBP + review system + social content + SEO + strategy" (6 items, no GEO/chatbot mentioned). GEO and Chatbot become the V2 upgrade story rather than V1 commitments.

The wedge claim's "AI chatbot" line in the business case becomes a "coming in V2" item rather than a V1 commitment. Worth surfacing in the next sales-marketing pass to align messaging.

Inclusion — 24/7 customer support

Workflow (locked Session 2):

  • Channel: in-app dashboard chat only (Q-SUP-1 = a). Keeps logs, history, and customer context in one place. Customer opens our app to get support — minor friction acceptable since the dashboard is already where they manage their services.
  • AI gate posture: Loose AI for V1 (Q-SUP-2). AI handles ~70% of conversations end-to-end; ~30% escalate to founder. Conservative launch posture — tightens to "Tight AI" (~95% AI / 5% escalate) once we have ~3 months of operational data showing the AI is reliable on common patterns.
  • Capped content edits stay inside Service 1 (Q-SUP-3 = a). 24/7 support covers chat / billing / troubleshooting / how-to questions. No edit-execution cost double-counted here.
  • Conversation volume assumption: ~5 conversations/customer/month average (light support customer = 1–2; heavy = 10+). Realistic for SMB tier.
  • Memory: customer's setup info preloaded into support agent's context per session. Vector store on existing VPS infra.

24/7 support cost-to-deliver per customer per month:

Component Cost / customer / month
AI tokens — Haiku conversations (3.5/mo, routine) ~€0.10
AI tokens — Sonnet conversations (1.5/mo, complex) ~€0.20
Founder escalation — 15 min/customer/mo (Loose AI posture) ~€5 (15 min @ €20/hr)
Vector store / context infrastructure share ~€0.10
In-app chat infrastructure share ~€0.10
Total 24/7 support inclusion ~€5.50/mo

Cost-to-deliver 24/7 support inclusion: ~€5.50/month per customer

Operational lever: Tight AI by month 3–6.

  • Loose AI (V1 launch): ~15 min/customer/mo founder time = €5/mo
  • Tight AI (proven by month 3–6): ~3 min/customer/mo founder time = €1/mo
  • Drops support cost from €5.50 to ~€1.50 once AI is trusted. That's the operational improvement that unlocks healthier margins, especially on à la carte and low-tier customers.

Capacity warning at scale: At 100 customers × 15 min/customer = 25 founder-hours/month on support alone. Combined with SEO Advanced founder review time (133 hr/mo at 100 Advanced customers), the founder-time ceiling triggers well before the cash math breaks. Both lines compress dramatically once Tight AI lands; that's the path that keeps the business solo-runnable past 100 customers.

Complete cost-to-deliver picture (V1)

Service Cost / customer / mo Notes
Website ~€3 Setup amortized + edits + hosting + escalation
GBP ~€2.50 OAuth, posts, review monitoring, photos
Review Engine ~€2.85 WhatsApp/SMS via owner's plan = €0 per send
Content (Mid tier, 16 posts/mo) ~€3.60 Postiz self-hosted, customer photo bank
SEO Basic (2 articles/mo, owner approves) ~€3.70 Free tracking path
SEO Advanced (4 articles/mo, founder review) ~€50.75 Seobility + founder time dominate
Strategy ~€0.75 Cheapest service, near-pure-margin €10 refresh upsell
24/7 support inclusion (loose AI, V1) ~€5.50 Per customer, not per service
24/7 support inclusion (tight AI, by mo 3–6) ~€1.50 Operational improvement

Reference customer scenarios (cost to us per customer per month):

Scenario Math Total cost-to-deliver
À la carte: Website only €3 + €5.50 support €8.50/mo
À la carte: Strategy only €0.75 + €5.50 support €6.25/mo
À la carte: Website + GBP + Reviews €8.35 + €5.50 support €13.85/mo
"Essential" bundle (everything Basic) €3 + €2.50 + €2.85 + €3.60 + €3.70 + €0.75 + €5.50 ~€21.90/mo
"Pro" bundle (SEO Advanced instead of Basic) €3 + €2.50 + €2.85 + €3.60 + €50.75 + €0.75 + €5.50 ~€68.95/mo
Same Pro bundle once Tight AI lands €3 + €2.50 + €2.85 + €3.60 + €50.75 + €0.75 + €1.50 ~€64.95/mo
Pro bundle with lighter SEO QA (month 6+) €3 + €2.50 + €2.85 + €3.60 + ~€31 + €0.75 + €1.50 ~€45.20/mo

Pricing structure — locked Session 2

Approach: Middle (FR-anchored). Boss locked 2026-05-07.

Service Retail (FR) Margin Notes
Website €29/mo 90% Anchor 1
GBP + Reviews (merged) €29/mo 82% Anchor 2
Content (mid tier, 16 posts/mo) €49/mo 93% Anchor 3
SEO Basic (Website add-on) €39/mo 91% Requires Website
Strategy (Website/Content add-on) €15/mo 95% Requires Website or Content
Strategy on-demand refresh €10 90% One-shot purchase
Essential bundle (all-in) €79/mo 72% All 5 services + Strategy + 24/7 support
Auto-mode upgrade (GBP+Reviews & Content auto-publish/auto-reply) +€19/mo 100% Pure margin lever
À la carte single-service setup fee €49 — Discourages margin-thin standalones
Custom-design upgrade €599 setup + monthly varies Real founder time involved
Annual prepay 2 months free — All plans
Founders pricing (first 50 customers) €59/mo Essential 63% Locked-for-life; closes after 50 sign-ups

Wedge realization at Middle:

  • Essential bundle €79/mo vs traditional agency retainer €500–1,500/mo → 6–19× cheaper
  • Wedge claim "at one-tenth the agency price" comfortably true at the agency-mid range
  • Margin at €79 = 72% gross, healthy for solo-founder operations

TN parity (when launch market locks): Divide retail by ~1.7× for TN-DT purchasing-power equivalence. Essential bundle would price ~140 DT/mo TN.

V1.5 move held in reserve: if Middle converts easily (>8% cold conversion), re-anchor to High pricing (€99 Essential), keep Middle as "Standard plan."

Skipped in V1: free trial, money-back guarantee. Same-day delivery is already the wedge — no need to add risk-removal incentives at launch. Add later only if conversion friction shows up in real data.

Marketing & customer acquisition

Reframe locked Session 2 (2026-05-08): the customer-facing AI is an onboarding agent, not a sales agent. Customers arrive already decided to subscribe (via paid ads, organic search, content, brand). Marketing does the convincing upstream; the agent runs interactive subscription + setup, Wix-style. See memory project_ai_agency_onboarding_agent.md.

Implication for go-to-market: standard SaaS marketing playbook applies (paid acquisition + content + SEO + social proof + landing-page CRO). Marketing is the bottleneck and the lever. Ads can fire from day 1 because we're not waiting for "AI sales tuning" — we're optimizing signup-flow completion, which is concrete and measurable in days.

Pre-launch marketing builds (one-time, not per-customer cost)

Build cost: €0 cash (Alfred/Claude Max plan covers production). Boss's time: ~hours of review + creative direction across these items, sized in marketing session.

  1. Landing page — high-converting hero + wedge claim + social proof slots (testimonials added post-first-cohort) + clear "Get Started" CTA + objection-handling copy + vertical-specific variants
  2. Brand presence assets — our own GBP for the AI Agency brand, LinkedIn page, basic social presence, brand-name domain + email
  3. Content seed — 5–10 vertical-specific landing pages targeting existing demand pools (e.g., "agence web Tunis", "agence SEO Paris", "refonte site internet", "site internet artisan"). Category-insertion content, not "AI Digital Agency" awareness content (the category is too new in our markets).
  4. Ad creative templates — video, image, copy variants for Meta + Google. Generate via AI image tools + Canva-style programmatic templates. ~3–5 base creatives per vertical at launch.
  5. Funnel analytics + heatmap — PostHog self-hosted on existing VPS (free), Microsoft Clarity (free) for heatmaps. Used for daily signup-flow optimization.
  6. Referral mechanic — single-link referral with €X credit on both sides, simple implementation. Captures word-of-mouth from first cohort.

Branding container + brand name: Boss's open question (Q10/Q11). Decided in Brand session, not here.

Marketing operations — ongoing cost lines

Line V1 launch (months 1–3) Scale (months 4–12)
Paid ad spend €500–1,000/mo €1,500–6,000/mo (scales with proven CAC)
Content production (AI-assisted, ongoing) ~€0 (token cost) ~€0
Funnel analytics + heatmap (PostHog, Clarity) €0 (free / self-hosted) €0
Ad creative iteration ~€0 (AI-generated) small budget for human-touched creative if needed
Brand presence maintenance €0 (Boss + AI handles) €0
Monthly marketing operations cost ~€500–1,000/mo ~€1,500–6,000/mo

CAC headroom math (Middle pricing, €79/mo Essential bundle):

  • LTV year 1 = €79 × 12 = €948
  • 3:1 LTV:CAC → max CAC ~€316
  • 5:1 LTV:CAC → max CAC ~€190
  • We can afford €150–300 paid CAC per Essential customer comfortably.

Preliminary 12-month customer-add curve (marketing-led)

Note: this curve is a placeholder pending refinement in the Marketing session. Boss has marketing ideas to brainstorm; the curve will be tightened against the channel-mix decided there.

Inputs: Middle pricing locked, marketing-led growth (not founder-outbound-throttled), €500–6K/mo ad spend ramped to proven CAC.

Phase Months Activity Customer adds/mo
Soft launch 1 Landing page live + €500 ads + brand presence + tighten signup flow 5–15
Founders pricing fill 2–3 €1,000/mo ads + organic content seeding takes hold + first testimonials 15–30
Public ads scaling 4–6 €1,500–3,000/mo ads, founders pricing closes after 50 sign-ups, testimonials power CRO 25–50
Full marketing engine 7–12 €3,000–6,000/mo ads, content marketing compounds organic, referral mechanic active 40–80

Cumulative trajectory:

Month Customers (cumulative, realistic case)
1 10
3 60
6 200
9 350
12 550

100 customers around month 4. This is meaningfully steeper than the founder-outbound-only curve (month 7–8), but driven entirely by marketing budget × conversion rate, not by founder hours.

Risks worth naming honestly:

  1. The curve assumes the marketing message + landing page convert. If conversion is weak, ad spend grows without commensurate customer adds. Mitigation: small budgets first, scale only on proven CAC.
  2. Categorical newness in our markets — we have to insert into existing demand pools (search for "agence web", "agence SEO"), not teach a new category. Shapes ad targeting and SEO content significantly.
  3. Brand-recognition gap at launch — must have basic brand presence (own GBP, LinkedIn, content seed) before first ad runs or trust collapses.
  4. Higher cash-flow risk in early months than founder-outbound — first 3 months total marketing spend ≈ €1,500–3,000 cash before significant revenue.

Customer-conversation requirement (mitigates the marketing-led learning gap)

Schedule explicit "talk to customers" loops for first 20–30 customers post-onboarding. Replaces the conversation-rich learning that sales-led outreach would have given us. Critical, not optional. Format: 15-min Zoom/WhatsApp call per customer at week 2 + week 6 + month 3 of their lifecycle. Document insights in a customer-discovery doc.

12-month cash projection scenarios

[pending — to be modeled against the marketing-led curve above and the locked Middle pricing]

Breakeven math

[pending — framed standalone against the business's own monthly cost structure (cost-to-deliver + marketing operations + tooling), not against personal burn. Modeled after marketing strategy session locks the marketing budget.]

Go / no-go thresholds

[pending]

Marketing Strategies

All Phase 1 / Phase 1.5 / Phase 2 strategies + reframed + rejected + formation-moment cluster.

Source: ~/claude-sync/business/products/ai-agency/marketing-strategies.md

AI Digital Agency — Marketing Strategies

Living doc. Umbrella for marketing channels and tactics. Cross-references partnerships-strategy.md for partner/referral channels specifically.

Each strategy gets locked here only after Boss has approved the concept. Active brainstorms are in conversation; this doc captures decided strategies.

📋 Master summary table — all strategies

Phase 1 — Launch-phase (fire from day 1, no capital required)

# Strategy Mechanism summary Status
A Reverse-Pitch Demos Pre-build complete demo deliverables for 50 high-fit prospects per cohort; outreach with "We built this for you, want it live?" ⭐ Locked
B Trade-Supplier Partnerships Partner with vertical distributors (L'Oréal Pro, Cedeo, MRS) for newsletter/catalog inclusion as recommended digital partner Locked
(n/a) Comptable Partnerships Cold outreach to small French comptables; commission-based referrals (in partnerships-strategy.md) Locked
D Educational Guide — digital 3 vertical-specific PDF guides as lead magnet + partner deliverable + social content + onboarding asset Locked
E Stamp (Tampon) Printer Partnerships Insert/email/checkout placement with small French stamp printers (Tampons Express, Tamponéo, IZI Tampon) ⭐ Locked, high-priority
F Business Card Printer Partnerships Same shape as E with smaller French card printers (Helloprint, Awisée, regional players) Locked, parallel to E
G Business Registration Cabinets Partner with small formation cabinets via packaging (checkout add-on) or advertising (post-registration emails) ⭐ Locked, most upstream
(n/a) Vertical FB Groups Post value-add content in commerce/professional Facebook groups (no pitching) Locked tactical presence
(n/a) Customer-as-Distributor Flywheel Unique referral code per customer (€20 credit) + co-branded QR/sticker checkout signage Locked, activates from customer #1
(n/a) Founder-led DM Outbound Curated 50-100 priority prospects, personalized DMs from Boss's personal account Locked light-volume tactic

Phase 1.5 — Revenue-gated (fire when MRR can fund)

# Strategy Triggers when Budget
C B2B Ad Campaign for Partner Recruitment First 5-10 partners closed manually €2,500-5,000 / 90 days
D' Educational Guide — physical mailing Digital opt-in conversion validated + ~€1,500 cash €1,250-2,000 for 500-piece test
— Reinvested ads (low-volume search/social) From month 3 onward, ~20-30% of prior MRR €100 → €500 → €1,500/mo

Reframed (concept rejected as marketing channel)

Concept Redirected to
Vertical booking platforms (Planity, Treatwell, Doctolib) Product integration (embed their booking widget in our sites)

Rejected

Concept Reason
Cold-comment outreach on prospect social media Meta TOS violation + brand-suicide + 0.1-0.5% conversion
CCI / BGE physical workshops Visa constraint (Boss can't enter France)

Phase 2 / Parking Lot (gated on ≥20-50 happy customers + recordable results)

Channel Gate
France Num activateur listing French legal entity + testimonials + results
Trade federations (UNEC, FNAA, CAPEB, CNAMS) Vertical case studies + member-discount structure
Pro accounting SaaS partnerships (Pennylane, Tiime, Indy) Customer-success data + integration roadmap
Channel arbitrage placements (paper boxes, PDF tools, receipt rolls) Revenue covers €200-500/test placements
Free utility tools we build Founder time available after primary channels validated
YouTube/IG SMB Audit content series Decision pending (brand-fronted vs founder-fronted)

The formation-moment cluster (defensibility argument)

Five Phase-1 partnership channels capture overlapping points in the new-SMB lifecycle:

Position Channel Mental state
1. Pre-registration Business registration cabinets (G) "How do I officially form this business?"
2. Just after SIRET Stamp printers (E) "I'm legally a business now"
3. First marketing assets Business card printers (F) "I need to look professional"
4. First operations Trade suppliers (B) "I need product to sell/use"
5. Ongoing Comptables "Who handles my accounting?"

Defensibility: a competitor showing up later has to negotiate with the same fragmented small-vendor landscape, partner by partner. Capital can't shortcut this work.

✅ Locked Strategies

Strategy A — Reverse-Pitch Demos ("Built it, want it?")

Locked Session 3, 2026-05-08.

Mechanic: Pre-build a complete demo deliverable for high-fit prospects BEFORE first contact. For each: their full new website (using their public photos / GBP info / IG content), optimized GBP draft, sample content batch (5 posts in their brand voice), SEO audit. Package into a personalized landing page (yourdemo.aiagency.fr/studiosophie).

Outreach format: "Bonjour Sophie, on a déjà construit ton nouveau site, ton GBP optimisé, et ton premier batch de contenu. Voici le résultat: [link]. Si ça te plaît, on le met en ligne ce soir pour €79/mo. Sinon, garde-le, c'est cadeau."

Why it works:

  • Demonstrates wedge directly — speed (already done) and quality (real output) before any € exchanged
  • Dissolves "trop beau pour être vrai" — output is right there, evaluable, undeniable
  • Asymmetric economics: at our cost-to-deliver (~€10/customer to build), 50 demos cost ~€500 total. At 10% conversion, 5 customers × €948/year LTV = €4,740 returned. ~9.5x experimental ROI.
  • Each demo doubles as a case-study asset (even rejections become testable proof for the next batch)
  • No competitor at our price tier can afford this — agencies' cost-to-deliver doesn't permit free demos. Our cost structure makes this uniquely possible.

Cost / founder-time:

  • ~30 min Boss-time per demo for review + send (build is automated by our pipeline)
  • 50 demos = ~25 founder-hours over 2-3 weeks
  • Marginal API cost: ~€1/demo = €50 total

Risks & mitigations:

  • Prospect takes the free site without subscribing — Mitigation: demo lives on yourdemo.aiagency.fr/... — they don't get the code/domain, GBP optimization requires us to push, content scheduling requires our pipeline. Theft requires effort that defeats the "free" appeal.
  • Demo feels stale if not converted in 7-14 days — Build expiration into the personalized landing page ("This demo expires on X")
  • Bad targeting wastes demos — Curate target list manually for first batch; refine criteria from results

This is also a content engine — every demo, even rejections, becomes social content ("On a construit le site de 50 salons en 3 semaines pour leur montrer ce qu'on fait. Voici les 10 plus belles transformations.")

Pre-launch builds required:

  • Personalized demo landing page system (template + per-prospect populator)
  • Pipeline that auto-builds the demo deliverables from prospect's public data
  • Outreach mechanic (which channel? email? IG DM? — TBD)
  • Tracking: who opened, who interacted, who converted
  • Curated target list (50 high-fit Sophie-shaped salons in target French cities)

Status: Concept locked. Build plan deferred to AI Agency PRD / architecture sessions.

Strategy B — Trade-Only Supplier Partnerships

Locked Session 3, 2026-05-08.

Mechanic: Trade-only distributors / suppliers ship catalogs and newsletters to tens of thousands of our exact ICPs:

  • Wellness/Beauty (ICP 1): L'Oréal Pro, Schwarzkopf Pro, Wella Pro
  • Trades (ICP 2): Cedeo, Brossette, Point.P (Saint-Gobain group)
  • Auto (ICP 3): MRS, IRP Auto, Mister-Auto Pro

These distributors don't sell websites — they sell physical product. Win-win partnership: they include us as their "recommended digital partner" in their newsletter / catalog / pro-customer welcome pack; we mention them in our content/templates as the supplier of choice. Mutual brand-association value, no zero-sum competition.

Why it works:

  • Massive distribution at zero CAC (a single distributor newsletter = 10k–50k pro accounts)
  • Trust transfer from a brand the merchant already trusts and pays
  • Recurring touchpoint (every product order = potential reminder)
  • Reciprocal value — they look more "modern / digital-friendly" by partnering with us
  • Fully remote partnership negotiation — visa-compatible

Risks:

  • Big distributors slow and bureaucratic; mid-size regional distributors may close faster
  • May require small partner fee or co-marketing budget contribution
  • Competitor (another digital provider) may already hold the slot — first-mover matters

First actions:

  • Build target list of 5 mid-size regional distributors per vertical (start with ICP 1)
  • Cold email + LinkedIn outreach to partnerships team
  • Pitch as "no cost to you, mutual brand value"

Cross-pollination with ReadyGo: Same model applies — ReadyGo can pilot supplier partnerships with Tunisian beauty/trades/auto distributors. Mirror approach to the comptable partnership pilot. Notion task TBD pending Boss decision on whether to fire ReadyGo-side first or in parallel.

Strategy C — B2B Ad Campaign for Partner Recruitment

Locked Session 3, 2026-05-08.

Concept: Run B2B paid ads marketing the partnership program itself (comptables + trade suppliers). Inverts the cold-outreach asymmetry — instead of hunting partners one-by-one, qualified partners self-select via ad → landing page → application.

Why it works:

  • Cold outreach: ~1-2 closes per 100 emails. Ads: 5k-50k targeted impressions per €100, with click-through filtering for intent.
  • LinkedIn job-title targeting is laser-precise for this audience (Expert-comptable, Partnerships Manager at L'Oréal Pro, etc.)
  • Self-selection delivers pre-qualified leads — clickers have signaled intent
  • Scales without founder-time bottleneck (ads run 24/7)
  • Campaign visibility itself positions us as established and scaleable
  • Compounds with manual outreach: first partners' testimonials fuel ad creative

When it fires — Phase 1.5 (NOT launch-day):

Phase Window Action
Phase 1 Now → +60 days Manual cold close of first 5-10 comptables + 5-10 suppliers (existing Notion tasks)
Phase 1.5 +60 → +120 days Launch B2B ad campaign — recruit partners at scale
Phase 2 +120 days+ Scale customer acquisition via self-feeding partner pipeline

Prerequisites before firing ads:

  • First 5-10 real partners closed manually (otherwise ads = vaporware optics)
  • Offer structure locked (commission rate + regulatory clearance for comptable-side)
  • Dedicated partner landing page (different from customer landing page)
  • Application → close pipeline ready (slow onboarding wastes ad spend)
  • First-cohort testimonials/numbers ready to use as creative

Channel mix (when fired):

  • LinkedIn (primary): B2B job-title targeting. CPC €5-15. €1,500-3,000 budget for first test.
  • Meta (secondary): Cheaper, broader targeting for owner-operators of smaller suppliers. €500-1,000.
  • Google Search (bottom-funnel): Capture intent keywords ("programme affiliation comptable," "partenariat digital salon"). €500-1,000.
  • Total Phase 1.5 test budget: ~€2,500-5,000 over 90 days.

Cross-pollination with ReadyGo: Same model can run for ReadyGo's partnership program. Cheaper Tunisian ad market (lower CPC) makes ReadyGo an even better staging ground for testing creative + offers before AI Agency budget commits.

Risks:

  • Without first manual partners, ads look like a shell game — sequencing matters
  • Offer structure must be locked first; "vague program" doesn't convert
  • Ad spend can be wasted if application flow leaks (onboarding ops gap)
  • LinkedIn CPC is brutal if targeting/creative misses

Status: Concept locked. Build plan deferred to Phase 1.5 trigger (after first 10 manual partner closes).

Strategy D — Educational Guide (Multi-Channel Asset + Direct Mail)

Locked Session 3, 2026-05-08.

Concept: Produce a printed/digital educational guide ("Le guide visibilité pour [vertical]") that teaches SMB owners why and how to be present online. The guide ends with our solution + QR code to the app. Physical mailing is one channel; the guide itself serves six channels as a single produced asset.

Why it works:

  • Cuts through digital noise (physical mail bypasses spam filters, algorithm gatekeepers)
  • Value-first positioning (educator, not pusher) — builds trust we don't have at launch
  • Reaches Marc (Trades) and Karim (Auto) personas who don't engage with digital marketing well but DO open their physical mail
  • Authority signal — printed material feels credible
  • Targeting precision via address-list filtering
  • Predictable cost (print + postage = known number)

The leverage move — one asset, six channels:

Use Channel Activation
1. Digital PDF download Lead magnet on landing page (email opt-in) Phase 1
2. Partner deliverable Comptables / suppliers hand to clients Phase 1
3. Social content Split into 10-15 FB / LinkedIn posts Phase 1
4. Onboarding agent asset Agent offers to email guide Phase 1
5. Physical mailing Direct mail campaign Phase 1.5+
6. SEO blog content Republish chapters as blog posts Phase 1-2

Vertical-specific (NOT universal):

  • "Le guide visibilité pour salons de coiffure et beauté" (ICP 1 — Wellness)
  • "Le guide visibilité pour artisans et services à domicile" (ICP 2 — Trades)
  • "Le guide visibilité pour garages et services auto" (ICP 3 — Auto)

A universal guide waters down the message. SMB owners want their own vertical in the examples.

Cost reality:

  • Print 12-16 page A5 booklet: ~€0.80-1.50/unit at scale
  • France postage (Lettre verte): ~€1.16/letter
  • Mailing house service (handles print + addressing + posting in France from Tunisia): +10-20%
  • Per-send delivered cost: €2.50-4
  • 500-piece test campaign: €1,250-2,000 budget

Conversion math (B2B direct mail typical):

  • Pessimistic 0.5% → CAC ~€500-800 (marginal)
  • Realistic 1.5% → CAC ~€188-200 (workable, ratio 3.5:1 against Year 1 LTV)
  • Optimistic 3% → CAC ~€83-100 (excellent, ratio 7:1+)

When it fires — sequenced phases:

Phase Window Action Budget Validates
Phase 1 m1-m3 Produce digital guides (3 vertical versions). Use as PDF lead magnet, partner deliverable, FB content. Test conversion on opt-in landing page. €0 (Boss + AI) Content converts
Phase 1.5 m4-m6 Print 200-500 copies. Pilot physical mailing to one curated vertical list. €500-1,500 (revenue-funded) Physical channel converts
Phase 2 m7+ Scale physical mailing if Phase 1.5 conversion holds. €3,000-12,000/month Pure scale

Critical sequencing rule: test digital before printing. If the digital version doesn't pull email opt-ins at 5%+, the physical version won't pull subscriptions.

Pre-launch builds required:

  • Guide content & structure (3 vertical versions, shared core + variant examples)
  • Professional design template (one-time)
  • PDF generation pipeline
  • Opt-in landing page with tracking
  • QR code + UTM tracking infrastructure
  • Address list sourcing (research vendors / scraping path)

Risks:

  • Production quality matters massively — cheap print kills credibility AND wastes postage
  • Address-list legality (GDPR considerations for B2B France — softer than personal but real)
  • Tracking attribution — without UTMs on QR codes, can't measure ROI per cohort
  • Bootstrap timing — physical mail can't fire at launch (€1,500+ budget exceeds m0 reserve)

Cross-pollination with ReadyGo: ReadyGo can use the same guide format for Tunisia market. Lower print + postage costs in Tunisia (~50% of France). Cheaper validation ground for the channel.

Status: Concept locked. Phase 1 (digital production) can start any time — €0 cash. Phase 1.5 (physical mailing) revenue-gated, expected m4-m6 in realistic scenario.

Strategy E — Company Stamp (Tampon) Printer Partnerships ⭐ HIGH-PRIORITY LAUNCH PLAY

Locked Session 3, 2026-05-08. Boss-identified during marketing brainstorm as "equivalent to comptable partnership, maybe even better."

Concept: Partner with French online stamp printers (Tampons Express, Tamponéo, IZI Tampon, Tampotex, Marquage.fr, etc.) to embed our offer at the exact moment a brand-new SMB receives their tampon — the most formation-locked physical purchase in French SMB lifecycle.

Why this beats both comptables and business card printers:

Property vs Comptables vs Business cards
Formation specificity Sharper — tampon requires SIRET (legal registration), business cards don't Sharper — many SMBs delay or skip business cards; nobody skips tampon
Universality Every SMB eventually has a comptable, but timing varies 100% of French SMBs need a tampon — legal/cultural requirement
Sales cycle SaaS-to-SaaS deal vs 21k human relationships Smaller, less-prestige printers — easier to negotiate than Vistaprint-tier
Buying-trigger moment Formation + ongoing Brand-new (decide name)

Mental state at the buying-trigger moment: The buyer just received their SIRET. They're committing publicly: "I'm officially in business now." Their mental frame is wide-open for "now build your digital presence" — peak receptivity.

Vendor landscape (French market):

  • Tampons Express (tamponsexpress.fr)
  • Tamponéo
  • IZI Tampon
  • Tampons Carrés
  • Tampotex
  • Marquage.fr
  • Stampie
  • Regional / smaller players (likely easier first closes)

Activation plays:

Touchpoint Mechanic
Delivery box insert Folded one-pager + QR code in every tampon shipment
Order confirmation email Sponsored placement: "Tampon prêt. Le prochain pas: votre site + Google Business Profile en 24h."
Checkout cross-sell Optional add-on at order: "Tampon + site web + GBP en bundle"
Co-branded landing page [partner].aiagency.fr/welcome with customer discount
Email follow-up sequence 7 days post-delivery: "Now that your business is officially official, here's what's next"

Cost / partnership currency:

  • Per-insert: €0.05-0.20 at volume
  • Order confirmation placement: typically flat monthly fee or revenue share
  • Smaller / regional printers: easier negotiation, lower threshold

Bootstrap fit:

  • First close requires only founder time (cold outreach + negotiation)
  • Pilot with one partner at €100-300 budget for inserts is revenue-fundable from m3-m4
  • Scales by closing additional printer partnerships sequentially

Cross-pollination with ReadyGo: Same model applies in Tunisia. Tunisian SMBs require a "cachet" (equivalent to French tampon). Smaller market = smaller printers = easier first closes. ReadyGo can pilot the channel with Tunisian stamp makers BEFORE AI Agency depends on it (same staging-ground pattern as comptables and suppliers).

Status: Concept locked. Build plan: Notion task for ReadyGo pilot (mirroring comptable + supplier task pattern).

Strategy F — Business Card Printer Partnerships

Locked Session 3, 2026-05-08. Parallel to Strategy E — same partnership shape, slightly different formation moment.

Concept: Partner with online business card printers (Vistaprint, MOO, OnlinePrinters, Exaprint, Helloprint, Awisée, regional French printers) to embed our offer at the moment new SMBs order their first cards.

Why parallel (not redundant) with Strategy E:

  • Stamps and business cards capture overlapping but distinct subsets of new SMBs
  • Different vendor landscape (no overlap between stamp printers and card printers)
  • Multiple touchpoints in formation moment = compounding capture rate
  • Same activation shape, can run in parallel without operational complexity

Why slightly weaker than Strategy E:

  • Business cards are not legally required (some SMBs skip them)
  • Card orders aren't gated by SIRET (some pre-formation orders happen)
  • Bigger printers (Vistaprint, MOO) require formal partnership teams — slower close. Use smaller / regional French printers for first deals.

Activation plays: Identical shape to Strategy E (insert, email upsell, checkout cross-sell, co-branded landing, follow-up sequence).

Vendor priorities: Start with mid-size French / European printers (Helloprint, OnlinePrinters, regional players), not Vistaprint-tier.

Status: Concept locked. Pursued in parallel with Strategy E once first stamp printer partnership closes.

Strategy G — Business Registration Services / Formation Cabinets ⭐ MOST UPSTREAM FORMATION CHANNEL

Locked Session 3, 2026-05-08. Boss-identified as next formation-moment channel — captures SMB owners at the most upstream point of all (BEFORE the SIRET is even issued, before stamps, before business cards, before comptables, before banks).

Concept: Partner with small business-registration cabinets and online formation services to either (a) package our offer as an add-on in their checkout flow, or (b) advertise our offer in their post-registration touchpoints. Customer arrives at us already deep in the "I'm setting up my business" mental state.

Why this might be the strongest formation-moment channel of all:

  • Most upstream moment — first commercial relationship a new SMB has. Their service ENDS where ours BEGINS (they finish registration → we begin digital presence).
  • 100% SMB audience — nobody uses formation services for fun.
  • Full profile captured — formation service has legal form, vertical, contact info, founder name.
  • Their customer is in "what else do I need?" mental state — peak receptivity to add-ons.
  • Mental flow is sequential and clear — registration → bank account → tampon → business cards → digital presence. We slot in after registration as logical next step.

Two activation modes

Mode 1 — Packaging (sharper, harder to negotiate)

Our service becomes a checkout add-on or bundled tier:

  • "Pack Création + Site Vitrine" — formation service charges €99 for registration + €X for our service (revenue share)
  • Or upsell at order completion: "Add a Site Vitrine + Google Business Profile in 24h for €79/mo?"
  • One click vs separate decision = highest conversion
  • Requires API/checkout integration → more setup work

Mode 2 — Advertising (faster to launch, lower conversion)

Sponsored placement in their existing flow:

  • Post-registration email sequence ("Vous êtes enregistré. Le prochain pas: votre présence digitale en 24h")
  • Customer portal banner / dashboard widget
  • Co-branded landing page with discount code for their customers
  • Welcome-kit insert for clients who get printed materials

Recommended: start with Mode 2 (advertising) for quick partnership closes, evolve to Mode 1 (packaging) with successful partners once trust is built.

Vendor strategy — small cabinets first

Tier Examples Why START with smaller
Big online platforms LegalStart, Captain Contrat, Pacte du Saint Pitch, LegalPlace Require formal partnership teams, credibility, case studies. Slow close. Save for Phase 2.
Small / regional cabinets ⭐ Independent business-formation lawyers, regional fiducies, solo entrepreneur-formation advisors, niche vertical-specific advisors Faster cycles, lower fees, partnership-revenue-hungry, no credibility gate. Start here.
Specialty-vertical formation services Services that focus on artisans (auto-entrepreneur), restaurants, etc. Best fit alignment with our ICP — vertical match doubles the relevance

Activation steps

  1. Build target list of 10-15 small French formation cabinets (regional, vertical-specialized, solo advisors)
  2. Cold outreach — email + LinkedIn + phone to founder/partner
  3. Pitch Mode 2 first (advertising/referral) — lower friction to close
  4. Negotiate first 2 partnerships — refine messaging based on feedback
  5. Build mechanic — co-branded landing page + tracking codes + welcome email template
  6. Materials package — partner-facing one-pager, FAQ, integration guide for portal embeds, email copy templates
  7. Evolve to Mode 1 (packaging) with high-performing partners once trust + conversion data exists

Cost / partnership currency

  • Affiliate commission per signup: 10-20% of first month or fixed €30-50 per converted customer
  • Sponsored placement in emails: typically €100-300/mo flat for smaller cabinets
  • Packaging revenue share (Mode 1): 20-30% of MRR for first 12 months — incentivizes the cabinet to push it

Bootstrap fit

  • Partnership negotiation = founder time only, near-€0 cost
  • First close requires only: pitch deck + co-branded landing page (€0 build under Claude Max)
  • Funded fully from day 1, no capital required

Cross-pollination with ReadyGo

Same model in Tunisia:

  • Tunisian cabinets de fiduciaires / fiduciary services
  • Small business-registration consultants
  • Tunisian equivalents to LegalStart (emerging local platforms)
  • Same logic: small > big, packaging > advertising once trust builds

ReadyGo pilot validates the playbook before AI Agency depends on it. Notion task created.

Status: Concept locked. Pursue in parallel with Strategies E (stamps) and F (business cards). Same project, same pattern, multiplying formation-moment surface area.

🔄 Reframed (concept rejected as channel, redirected)

Vertical Booking Platforms (Planity, Treatwell, Doctolib, Allotrade)

Reframed Session 3, 2026-05-08.

Why not a marketing channel: These platforms sell hosted booking pages to merchants — which is partial competition with our website service. They have zero incentive to refer their merchants to us, because we'd unbundle one of their products.

Where it goes instead — product integration: Embed THEIR booking widget into OUR customer sites as a product feature. Logic flips:

  • We become a distribution channel for them (every salon we onboard using their widget = more bookings flowing to their platform)
  • Salon customer gets real booking integration without us building one
  • Marketing claim becomes: "Booking Planity intégré gratuitement" (or Doctolib for healthcare adjacents, Treatwell for beauty, etc.)

Marketing-channel value comes later (Phase 2). Once we've moved 50–100 salons all running Planity widget, we have leverage to negotiate co-marketing or newsletter inclusion — because at that point we drive them merchant signups, not the reverse.

Action: Promote Planity widget integration to the AI Agency PRD / architecture session as a product feature, not marketed as a partnership at launch.

⏸️ Phase 2 Parking Lot (gated on proof)

These channels need us to have shipped + accumulated case studies + customer success stories before they'll close. Worth pursuing once we have ≥20-50 happy customers and recordable results.

Channel Gate (what we need before pursuing)
France Num activateur listing Need French legal entity (probably required) + customer testimonials + measurable results to qualify
Trade federations (UNEC, FNAA, CAPEB, CNAMS) Need vertical-specific case studies the federation can vouch for; member-discount deal structure
Pro accounting SaaS (Pennylane, Tiime, Indy) Need customer-success data + integration roadmap to make the partnership pitch credible

These all promise high leverage. They just don't fire at launch. Schedule revisit ~6 months post-launch (or when first 20 customers are happy and quotable).

🔍 Active Brainstorm (Session 3 ongoing)

Tracked in conversation. Items reaching consensus get promoted to "Locked" above.

Brainstorm directions still open:

  • Other remote-deliverable channels (content, paid, product-led)
  • Customer-as-distributor flywheel (referral codes + co-branded checkout signage from Idea 3 earlier)
  • Free SMB Audit content series (from Idea 2 earlier — pending decision on brand-fronted vs founder-fronted)

❌ Considered & Rejected

Cold-comment outreach on prospect social media

Rejected Session 3, 2026-05-08.

Why rejected:

  • Spam under Meta Community Standards — IG/FB account gets shadowbanned then banned (we'd lose @brand before closing 10 customers, and IG IS our primary channel for ICP 1)
  • Brand-suicide at the worst audience: prospects' followers ARE other prospects; one spammy comment = 50 prospects permanently lost
  • Inverts the wedge: positions us as "scrappy / desperate / cuts corners" instead of "professional / premium / AI-quality"
  • Conversion math brutal: 0.1-0.5% rate, 50-200 founder-hours for 10 customers, AND brand damage
  • No scalable "good" version: manual = grueling, automated = explicit bot-spam → instant ban

The directional instinct (be present where prospects are) was correct. Redirected to vertical FB groups (post value-add content, never pitch) and curated founder-led DM outbound from Boss's personal account.

CCI / BGE physical workshops

Rejected Session 3, 2026-05-08.

Why rejected: Boss has no Schengen visa, can't physically enter France. The directional instinct (institutional trust transfer to new business owners) is good — being explored via remote-only equivalents (online webinars, content partnerships, France Num listing).

Cross-doc dependencies

  • customer-profiles.md — defines who each strategy targets
  • customer-personas.md — calibrates voice/copy/channel for each strategy
  • scope-boundaries.md — every claim in marketing copy must trace here
  • partnerships-strategy.md — partnership-specific channels (comptables locked)
  • business-case.md + lean-plan.md — pricing math and wedge claims must stay consistent

Partnerships Strategy

Comptable (accountant) partnership channel — decisions, open items, ReadyGo cross-pollination.

Source: ~/claude-sync/business/products/ai-agency/partnerships-strategy.md

AI Digital Agency — Partnerships Strategy

Living doc. Captures decisions on partnership-channel marketing, separate from paid acquisition and content marketing.

Channel 1 — Accountant (Comptable) Partnerships

Hypothesis

Comptables meet new business owners at the moment of company formation — exactly when "do I need a website?" is unanswered. They're already trusted advisors, dissolving our dominant "trop beau pour être vrai" objection on contact. CAC collapses (referral fee instead of ad spend), leads are pre-qualified, and the channel reaches ICPs 2 and 3 (Marc, Karim) better than IG ads ever will.

Decisions locked Session 3, 2026-05-08

Question Decision
Offer to comptable Commission — contingent on legality under Ordre des Experts-Comptables rules
Regulatory boundary ⚠️ Pending — Boss to check the deontological code
Pilot strategy Cold outreach to target list (no warm network)
Mechanic Unique referral code per comptable; comptable sets their own discount for their clients; comptable can see results (reporting format TBD)
Vetting Customer-fit rules applied at the customer side, not the partner side. Any comptable can refer; ICP filter triggers when the client lands.
Comptable's post-referral role Hands-off after referral

Open items (must resolve before pilot)

  1. Regulatory check — Read Ordre des Experts-Comptables deontological code or consult one comptable directly. Confirms whether direct commission is legal, and if not, what alternative structures are (preferred-provider partnership, free service for their firm, co-marketing, white-label).
  2. Commission rate — Fixed across all partners or negotiable per comptable? Recurring (e.g., 10–20% MRR for 12 months) vs. one-time finder's fee? Lock the structure.
  3. Mechanic build — Unique referral codes are simple. The "comptable can see results" piece needs a build:
    • MVP: monthly email report per comptable
    • V2: lightweight dashboard (referrals, conversions, MRR earned, payout history)
  4. Pilot target list — Source list of French comptables we'd cold-outreach. Options: scrape Ordre's public directory, buy a B2B lead list, manual curation by region.
  5. Materials package — One-pager for the comptable explaining the offer; comptable-facing FAQ; copy-paste referral message templates (email + WhatsApp + in-person script); landing page parameter that auto-applies the comptable's discount.
  6. Vetting safeguards — Even with hands-off model, we need to monitor referral quality per comptable. If one comptable's referrals churn at 3× the average, something's wrong (mis-qualified clients or bad-faith referral). Build the metric early.

Why this works for ICPs 2 and 3 specifically

  • Marc (Trades, plumber): Ignores IG. Reads FB. Trusts maybe three professional voices in his life — his comptable is one of them. Cold ads convert him at near-zero.
  • Karim (Auto repair): "J'ai déjà été arnaqué par une agence." His comptable saying "c'est sérieux" eliminates the trauma block.
  • Sophie (Wellness): Comptable referral works but IG ads also work for her — comptable channel is supportive here, not primary.

Cross-pollination with ReadyGo

The same partnership model is being piloted on ReadyGo (the Tunisia website business) with local Tunisian comptables — Notion task created. Tunisia pilot will surface lessons (regulatory differences, comptable receptivity, mechanic friction, conversion rates) BEFORE AI Agency needs to depend on the channel. Use ReadyGo as the cheap learning ground.

Status

  • Idea locked, decisions captured
  • Execution gated on: AI Agency launch readiness + regulatory check + ReadyGo pilot learnings
  • Not in active build yet

6-Month Projections

Bootstrap-mode scenarios (pessimistic / realistic / optimistic) + month 7-12 realistic extension.

Source: ~/claude-sync/business/products/ai-agency/projections.md

AI Digital Agency — 6-Month Bootstrap Projections

Rebuilt Session 3, 2026-05-08 after critical reframe: bootstrap-mode is a hard constraint (Boss in debt, no launch capital available). All projections funded by founder time + first-customer revenue. See memory project_boss_bootstrap_only_mode.md.

Validation horizon: 6 months.

The reframe

Earlier model (rejected): €10k operating capital deployed over 6 months, breakeven at month 8-9.

Why it was wrong: Boss has no €10k. The default-SaaS-marketing-budget assumption (€1,500-3,000/mo paid acquisition) was bolted on top of strategies that don't actually need ads. The locked strategies are founder-time-driven; ads were decoration, not load-bearing.

The corrected model — bootstrap mode:

  • Month 0 cash gate: ≤ €100-200 (domain, first month basic tooling)
  • Months 1-2: zero paid acquisition. Reverse-pitch demos + cold partner outreach + organic + Founders pricing FOMO.
  • Month 3+: small reinvested ads (~20-30% of MRR) only if revenue covers them
  • Phase 1.5 B2B partner-recruitment ads: revenue-gated, not calendar-gated. Fire when MRR can fund.
  • No outside capital injection required. The business pays for itself from customer #1.

This works because our locked strategies were already designed for this:

  • ✅ Reverse-pitch demos — founder time + ~€1/demo API cost (trivial)
  • ✅ Comptable partnerships — cold outreach = founder time
  • ✅ Supplier partnerships — cold outreach = founder time
  • ✅ Customer-as-distributor flywheel — costs paid by customer revenue
  • ✅ Vertical FB group content — pure founder time
  • ✅ Founders pricing FOMO — built into pricing, costs €0 to activate

Locked input parameters

Parameter Value Source
Founder time available 78 hr/month (3 hr/day × 26 days) Boss Session 3
Validation horizon 6 months Boss Session 3
Capital available ~€100-200 m0 setup only Boss Session 3
Cost-to-deliver Essential bundle €21.90/customer/mo lean-plan.md (locked)
Cost-to-deliver à la carte avg ~€10/customer/mo lean-plan.md (locked)
Tooling fixed (no Seobility — V2) ~€85/mo Lean tooling stack
Founders pricing First 50 customers €59/mo locked lean-plan.md
Standard bundle pricing €79/mo after Founders cap lean-plan.md
Bundle vs à la carte mix (assumption) 60/40 Hypothesis to validate
Blended ARPU (Founders era) ~€53/customer/mo 60%×€59 + 40%×€45
Blended ARPU (Standard era) ~€65/customer/mo 60%×€79 + 40%×€45
Monthly churn (assumption) 2% Hypothesis to validate

Tooling (fixed costs)

Without Seobility (deferred to V2 with SEO Advanced):

Item €/month
Marketing site hosting + domain €20
Email service (transactional) €20
Domain renewals + SSL €5
CRM / pipeline (free tiers) €0
Postiz self-hosted on existing prod VPS €0
Cloudflare Turnstile + analytics €0
PostHog / Microsoft Clarity (free tiers) €0
Buffer for unexpected €40
Fixed tooling total ~€85/month

Three bootstrap scenarios

Pessimistic — channels underperform, slow trickle

Assumptions:

  • Reverse-pitch demos: 30/month at 3% conversion = 1 customer/month
  • Partner closes: 0 by month 3, 1 by month 6
  • Customers per partner: 0.5/month
  • Founders FOMO: minimal effect
  • Reinvested ads: €0 throughout (no buffer)
  • Churn: 5% monthly
Mo New Active MRR Cost-deliver Tooling Ads Total cost P&L Cumulative cash
0 — 0 €0 €0 €0 (setup) €0 €100 setup -€100 -€100
1 1 1 €53 €17 €85 €0 €102 -€49 -€149
2 2 3 €159 €51 €85 €0 €136 +€23 -€126
3 3 6 €318 €102 €85 €0 €187 +€131 +€5
4 4 10 €530 €170 €85 €0 €255 +€275 +€280
5 5 15 €795 €255 €85 €0 €340 +€455 +€735
6 6 21 €1,113 €357 €85 €0 €442 +€671 +€1,406
  • 6-month customers: 21 active
  • 6-month MRR: €1,113
  • 6-month cumulative cash position: +€1,406
  • Verdict at 6-month gate: Slow but cash-positive. Decision: persist with patience or pivot the customer ramp.

Realistic — channels work, modest reinvestment

Assumptions:

  • Reverse-pitch demos: 50/month at 8% conversion = 4 customers/month
  • Partner closes: 1 by month 3, 3 by month 6
  • Customers per partner: 2/month
  • Founders FOMO: 30% pull-forward effect months 1-3
  • Reinvested ads: 0 / 0 / 100 / 200 / 350 / 500 (~20% of prior month MRR, gated on cash)
  • Churn: 2% monthly
Mo New Active MRR Cost-deliver Tooling Ads Total cost P&L Cumulative cash
0 — 0 €0 €0 €0 €0 €100 setup -€100 -€100
1 4 4 €212 €68 €85 €0 €153 +€59 -€41
2 6 10 €530 €170 €85 €0 €255 +€275 +€234
3 8 18 €954 €306 €85 €100 €491 +€463 +€697
4 10 27 €1,431 €459 €85 €200 €744 +€687 +€1,384
5 13 39 €2,100 €663 €85 €350 €1,098 +€1,002 +€2,386
6 15 53 (Founders cap hit ~m5) €2,902 €901 €85 €500 €1,486 +€1,416 +€3,802
  • 6-month customers: 53 active (Founders cap hit around m5)
  • 6-month MRR: €2,902
  • 6-month cumulative cash: +€3,802
  • Verdict at 6-month gate: Healthy. Business pays for itself, builds €3.8k cash reserve, can fund Phase 1.5 partner-recruitment ads from m7+.

Optimistic — channels overdeliver, aggressive reinvestment

Assumptions:

  • Reverse-pitch demos: 60/month at 15% conversion = 9 customers/month
  • Partner closes: 3 by month 3, 6 by month 6
  • Customers per partner: 4/month
  • Founders FOMO: 60% pull-forward, sells out by month 4
  • Reinvested ads: 0 / 100 / 300 / 600 / 1,000 / 1,500 (~30% of MRR, aggressive)
  • Churn: 1% monthly
Mo New Active MRR Cost-deliver Tooling Ads Total cost P&L Cumulative cash
0 — 0 €0 €0 €0 €0 €100 setup -€100 -€100
1 8 8 €427 €136 €85 €0 €221 +€206 +€106
2 14 22 €1,175 €374 €85 €100 €559 +€616 +€722
3 22 44 €2,350 €748 €85 €300 €1,133 +€1,217 +€1,939
4 24 67 (Founders sold out) €4,015 €1,139 €85 €600 €1,824 +€2,191 +€4,130
5 28 94 €5,946 €1,598 €85 €1,000 €2,683 +€3,263 +€7,393
6 32 124 €8,094 €2,108 €85 €1,500 €3,693 +€4,401 +€11,794
  • 6-month customers: 124 active
  • 6-month MRR: €8,094
  • 6-month cumulative cash: +€11,794
  • Verdict at 6-month gate: Strong scale. €11.8k cash on hand to fund Phase 1.5 ads, first hire (~€2.5k/mo Tunisia-based junior support), or both.

Decision-grade outputs

Capital gate

Scenario M0 setup needed Cash position at any point Verdict
Pessimistic ~€100 Briefly negative m1 (-€149), positive from m3 Tight but survivable
Realistic ~€100 Briefly negative m1 (-€41), positive from m2 Healthy
Optimistic ~€100 Cash-positive from m1 (+€106) Excellent

The capital gate is ~€100-200 for m0 setup. That's the entire launch capital required. After that, the business funds itself.

If even €100 is genuinely unavailable, mitigation options:

  • Pre-sell 1-2 Founders subscriptions before m1 (collect €59 each = covers setup)
  • Small ReadyGo profit injection (cross-subsidy from existing business)
  • Defer launch by 1-2 months while saving €200 specifically for this

Cash flow danger zone

Realistic scenario: Briefly negative cash in early months (-€149 at end of m1, recovering by m3).

This is the cash flow lag risk: customers paying monthly upfront still leaves a 1-month gap before revenue covers cost. Need to ensure m0 setup + m1 minimum cash reserve = ~€200-300 to avoid running negative.

Operating breakeven (revenue covers monthly costs)

Scenario Operating breakeven
Pessimistic Month 2
Realistic Month 1
Optimistic Month 1

All scenarios hit operating breakeven within 2 months — the business is structurally profitable from very early on. The question isn't "when do we break even" but "how fast does the customer base grow."

Founder-time check

Realistic scenario at m6:

  • Demos: 25 hr/mo (50 demos × 30 min)
  • Partner outreach: 10-15 hr/mo (winding down as Phase 1.5 ads activate)
  • Customer ops: ~13 hr/mo (53 customers × 15 min @ Loose AI)
  • Misc / strategy / content: 10 hr/mo
  • Total: ~63 hr/mo of 78 available

Holds through 6 months. Bottleneck likely m7-8 in realistic, m4-5 in optimistic.

Phase 1.5 B2B partner-recruitment ads (revenue-gated)

Scenario When MRR covers €500/mo budget When Phase 1.5 fires
Pessimistic Not within 6-mo horizon Defer beyond m6
Realistic Month 4-5 Month 5-6 (small budget, scale m7+)
Optimistic Month 3 Month 3-4 (full budget)

Key sensitivities (what can swing the curve)

The few variables that matter most — track these against reality and lock the scenario after 8 weeks of data:

  1. Reverse-pitch conversion rate (3% / 8% / 15%) — biggest swing variable. First 50 demos = the most important data we'll ever collect.
  2. Customers per partner per month (0.5 / 2 / 4) — partner productivity is highly uncertain at zero proof.
  3. Churn rate (1% / 2% / 5%) — small early effect, dominates LTV later.
  4. Founders FOMO pull-forward (small / 30% / 60%) — front-loads validation signal.
  5. Bundle vs à la carte mix (40/60 / 60/40 / 75/25) — affects ARPU and cost-to-deliver mix.

What this projection does NOT include (intentionally)

  • Brand identity investment (DIY at launch; polish later when revenue allows)
  • French legal entity for France Num (Phase 2; revenue-gated)
  • Founder salary (€0 cash — Boss is in debt, not pulling salary)
  • Tax / accounting fees (deferred to legal-structure decision)
  • Personal expenses comparison (per project_ai_agency_standalone_framing.md)

Decision summary

Capital gate: ~€100-200 for m0 setup. Match it, launch is on the table.

Operating profile:

  • All three scenarios cash-positive within 6 months
  • Realistic scenario builds €3.8k cash reserve, enables Phase 1.5 ads from m7
  • Pessimistic still cash-positive (slow trickle), produces validation signal at m6
  • Optimistic compounds aggressively, enables first hire by m6-7

Critical insight from the rebuild: the strategies we already locked (reverse-pitch, partnerships, customer-as-distributor) are bootstrap-native. The "we need €10k for ads" framing was a default-SaaS habit, not load-bearing on our actual go-to-market. Removing it doesn't compress the curve much — it removes phantom spend.

The 6-month decision at the validation gate:

  • Cash + customer count on pessimistic curve → review whether to pivot the channel mix or accept slow growth
  • Cash + customer count on realistic curve → continue, fund Phase 1.5 ads from accumulated cash
  • Cash + customer count on optimistic curve → scale aggressively, plan first hire, lock V2 timeline

Realistic scenario — months 7-12 extension

Continuation of the realistic scenario beyond the validation gate. Shows what success looks like if month 6 lands on the realistic curve.

Continuing assumptions:

  • Reverse-pitch demos: 50/month × 8% = 4 customers/month
  • Partners: 4 (m7) → 5 (m9) → 6 (m11) → 7 (m12)
  • Customers per partner: 2/month
  • Phase 1.5 B2B partner-recruitment ads activate from m7 (revenue covers)
  • All new customers from m7 onward are Standard pricing (€79 bundle)
  • First hire (Tunisia junior support, ~€2,500/mo) starts mid-m9
  • Tooling rises to €120/mo from m9 (scaling buffer)
  • Churn: 2%/mo continues
Mo New Churn Active Founders/Std ARPU MRR Cost-deliv Tooling Ads Hire Total cost P&L Cum. cash
6 15 1 54 48/6 €54.73 €2,956 €945 €85 €500 — €1,530 +€1,426 +€3,841
7 16 1 69 47/22 €57.23 €3,949 €1,208 €85 €700 — €1,993 +€1,956 +€5,797
8 18 1 86 46/40 €58.98 €5,072 €1,505 €85 €1,000 — €2,590 +€2,482 +€8,279
9 22 2 107 45/62 €60.36 €6,459 €1,873 €120 €1,500 €1,250 €4,743 +€1,716 +€9,995
10 24 2 129 44/85 €61.31 €7,909 €2,258 €120 €2,000 €2,500 €6,878 +€1,031 +€11,026
11 25 3 152 43/109 €62.01 €9,425 €2,660 €120 €2,500 €2,500 €7,780 +€1,645 +€12,671
12 28 3 177 42/135 €62.55 €11,071 €3,098 €120 €3,000 €2,500 €8,718 +€2,353 +€15,024

12-month realistic outcome

  • Active customers: 177
  • MRR end of month 12: €11,071
  • ARR (Annualized Recurring Revenue) run-rate: €132,852
  • Cumulative cash: +€15,024
  • Team: Boss + 1 junior hire (Tunisia, ~€2,500/mo)
  • Phase 1.5 partner ads steady at €3,000/mo budget, fully self-funded

Key milestones (m7-m12)

  • M7: Phase 1.5 B2B partner-recruitment ads activate (revenue-funded)
  • M8: MRR crosses €5,000 — hire becomes affordable
  • M9: First hire starts mid-month (half-cost month)
  • M10: Hire fully productive, founder-time freed
  • M11-12: Steady-state growth, V2 services upsell potential to existing customers

Founder-time at month 12

Activity Hours/month
Reverse-pitch demos 25
Partner outreach (mostly ads-driven now) 5
Customer ops (15 min × 177 = 44 hr at Loose AI; halved by hire) 22
Misc / strategy / content / escalations 10
Total founder hours ~62 of 78 available

If "Tight AI" support posture lands by m6 (per lean-plan §Support, drops support to 3 min/customer/month), founder hours could drop to ~30-40, freeing capacity for V2 build and strategic work.

What's NOT in the 12-month extension

  • V2 launch (AI Chatbot, GEO, TikTok, AI image gen, SEO Advanced) — slots in around m9-12 typically
  • Second hire (content/SEO specialist or partnerships manager) — likely m12-15
  • Geographic expansion (Tunisia / English markets) — deferred until France core proven
  • External funding — none assumed; bootstrap continues
Snapshot of ~/claude-sync/business/products/ai-agency/. Pre-launch, brand name not yet locked. This page is temporary and will be removed.