diff --git a/docs/workshop/bmc.md b/docs/workshop/bmc.md
new file mode 100644
index 0000000..95b7f76
--- /dev/null
+++ b/docs/workshop/bmc.md
@@ -0,0 +1,89 @@
+# Business Model Canvas — travelledger
+
+> Converged canvas from a double-diamond workshop (diverge → converge).
+> Five role-based sub-agents produced divergent options (3 strategic directions,
+> 6 personas, 4 revenue variants, 5 stacks). This is the *converged* view, with
+> the strongest divergent alternatives noted inline. All numbers are illustrative
+> assumptions to be validated.
+
+## Positioning decision (converged)
+**Launch as Direction A (neutral travel ledger), sequence into Direction B (booking
+commerce).** Rationale: lowest barrier to entry, no partner dependency at launch,
+validates the core habit (will users log travel data?) before monetization pressure.
+Commissions (Direction B) have 10x revenue potential but need trust + scale → earn it
+as a neutral tool first, then introduce booking as convenience, not extraction.
+
+---
+
+## 1. Customer Segments
+- **Frequent leisure travelers** — track budget, preserve memories, avoid budget disasters.
+- **Business travelers / consultants** — reimbursements, separate personal vs. business, cancellation deadlines. (High retention if workflow-integrated.)
+- **Family trip organizers** — coordinate multi-person itineraries, split costs, store confirmations.
+- **Digital nomads** — visa/tax-day tracking, multi-currency, long-term narrative.
+- **Travel agents (B2B, non-obvious)** — manage client itineraries, monitor deadlines across suppliers.
+- *(Deferred)* SMB corporate travel managers (Direction C) — revisit if B2C shows organic business adoption.
+
+## 2. Value Propositions
+- **One calm place** for the whole trip: flights, hotels, trains, expenses, QR codes.
+- **Cancellation Guardian**: proactive alerts before deadlines (the defensible, emotional pain point).
+- **Post-trip statistics**: where the money actually went, across currencies.
+- **AI assist (later)**: auto-categorize expenses, parse booking emails/PDFs into itinerary.
+- **Purchase opportunities (later)**: book hotels/tours in-app with affiliate convenience.
+
+## 3. Channels
+- SEO ("travel budget template", "expense tracker for trips"), referral ("invite a travel buddy"),
+  micro-influencers (nomad YouTubers, family-travel Instagram), OTA post-purchase partnerships
+  ("Save this booking to travelledger"), Reddit/expat/FB groups, later app stores (iOS/Android).
+- **Viral loop:** shared trip page → non-users see a beautiful itinerary with QR + expense summary
+  → "get alerts / add your expenses free" → they invite their network.
+
+## 4. Customer Relationships
+- **Self-serve freemium** (table stakes) + **AI concierge "Trip Guardian"** (the moat: proactive
+  deadline alerts, auto-categorization). B2B account management deferred for travel-agent segment.
+
+## 5. Revenue Streams
+| Model | Price (illustrative) | Role |
+|-------|----------------------|------|
+| Freemium subscription | Free: 1–5 trips/yr; Premium $4.99/mo or $39/yr; Pro $9.99/mo (family/multi-user) | **Primary (≈60%)** |
+| Per-trip pass | $2.99–$5.99/trip | Secondary (≈25%) — captures episodic users |
+| Booking commission | Hotels 3–8%, tours 8–12%, insurance 15–25%, flights 1–3% | Tertiary upside (≈15%) |
+| B2B licensing (deferred) | $8–15/user/mo; agent white-label $30–50/mo | Post-PMF |
+
+## 6. Key Resources
+- Mobile-first web app (Next.js + Capacitor), cloud backend (Supabase/Firebase),
+  encrypted attachment storage (R2/S3), OCR + LLM for parsing/categorization,
+  push infrastructure for deadline alerts.
+
+## 7. Key Activities
+- App + AI development, user acquisition (SEO/content), partner integration (OTAs,
+  banks via Open Banking), compliance (GDPR, affiliate disclosure), data/analytics.
+
+## 8. Key Partners
+- **OTA aggregators** (Booking.com, Expedia, Skyscanner) — inventory + affiliate tracking.
+- **Hotels / tour ops** (GetYourGuide, Viator) — direct rates, higher-margin commissions.
+- **Airlines / rail** — data, booking links, status.
+- **AI/LLM providers** — parsing + categorization (provider-agnostic interface).
+- **Banks / cards** (Revolut, Wise, Plaid/PSD2) — auto expense capture.
+- **Travel-insurance** providers — high-margin embedded quotes.
+
+## 9. Cost Structure
+| Category | Bootstrapped (early) | Funded (scale) |
+|----------|----------------------|----------------|
+| Cloud/hosting | $200–800/mo | $5k–20k/mo |
+| Storage (QR/images) | $50–200/mo | $1k–5k/mo |
+| AI/LLM inference | $100–500/mo | $10k–50k/mo |
+| Dev (founder+contractors) | $5k–15k/mo | 8–12 FTEs |
+| Support / compliance | ~$0–2k / ~$500 | $15k–25k / $10k–30k |
+| CAC | SEO $5–15, paid $25–60, partner $10–40 | similar, higher volume |
+
+## Key risks (cross-role)
+- **Churn / seasonality** — travel is episodic → annual plans, trip passes, reserves.
+- **Commission clawbacks & API deprecation** — diversify partners; no partner >25% of revenue.
+- **Chargeback exposure** — use affiliate networks (they are merchant of record) early; avoid MoR status.
+- **Commission cash-flow lag (30–90 days)** — keep 3-month runway; don't count as immediate revenue.
+- **QR extraction variance** across vendors — LLM parsing, not regex.
+
+## Day-one metrics
+1. **Paid conversion rate** (target 3–5% of MAU)
+2. **Trip completion rate** (>60% of started trips)
+3. **CAC payback period** (<6 months)