mancitrus/system-prompts-and-models-of-ai-tools

mirror of https://github.com/x1xhlol/system-prompts-and-models-of-ai-tools.git synced 2026-06-18 15:29:36 +00:00

Business Viability Kit: discovery-phase operating artifacts

Saves the DEALIX_BUSINESS_VIABILITY_KIT.md (Weeks 4-12 customer discovery
operating manual) and produces only the operational artifacts it explicitly
names. Per the kit's Appendix C: no new plan documents, no Wave A-E work,
no features without customer pull.

Added:
  Customer Viability operating artifacts
    - docs/customer_learnings/hypotheses.yaml - 12 hypotheses tracked
      to SUPPORTED/FALSIFIED/AMBIGUOUS with interview-log citations
    - docs/customer_learnings/interviews/_template_ar.md - 45-min Arabic
      discovery script + post-call log schema
    - docs/customer_learnings/interviews/_template_en.md - English version
    - docs/customer_learnings/founder_dashboard.md - weekly Monday printable
      dashboard (kit Sec 8)
    - docs/customer_learnings/pricing_discovery.md - Van Westendorp PSM +
      value-based sanity check + A/B model matrix
    - docs/customer_learnings/unit_economics.md - per-customer economics,
      LTV/CAC ratios, 12-month scenario template
    - docs/customer_learnings/defensibility_scorecard.md - 5 moats x 2
      questions, quarterly re-measurement

  Registry updates
    - docs/registry/TRUTH.yaml customer_validation section: hypothesis
      counters + discovery-interview counter + kit reference
    - docs/customer_learnings/README.md updated to link new artifacts

Gates after change:
  architecture_brief.py       40/40
  release_readiness_matrix  102/102  (added 8 new BVK artifact checks)
  v005_truth_registry_audit   19/19 SUPPORTED

Agent scope going forward per kit Appendix C: customer-surfaced P0 defects,
UX polish appearing in 2+ interviews, perf issues on staging, pentest
remediations. No new plans. No Wave tasks.

2026-04-17 11:26:32 +00:00

3.5 KiB

Raw Blame History

Dealix — Unit Economics Worksheet

Fill in ONLY after 3 paying customers. Filling it earlier is fiction. All figures in SAR unless stated. Re-run at end of each month once populated.

1. Per-customer monthly economics

Revenue

MRR per customer (avg): ______

Cost to serve (per customer / month)

Line	Amount (SAR)
Infrastructure (AWS + DB + Redis)	______
LLM API (Anthropic + OpenAI + Groq, at observed usage)	______
Third-party services (Sentry, OTel backend, etc.)	______
CS time (hours × hourly cost)	______
Support time (eng on-call, tickets)	______
Total Cost to Serve	______

Gross Margin

GM per customer (SAR): ______
GM %: ______ %
Target: GM ≥ 70% (SaaS healthy)
Red flag: GM < 60% → LLM spend dominates; audit model routing.

2. Customer Acquisition Cost (CAC)

Line	Amount (SAR)
Founder time spent closing (hrs × opportunity cost)	______
Marketing spend (ads, events)	______
Sales tools (CRM, LinkedIn Sales Nav, etc.)	______
Referral incentives	______
Total CAC (amortized across paid customers)	______

3. Lifetime Value (LTV)

Expected contract length (months, honest not aspirational): ______
Monthly gross margin per customer: ______
LTV = months × monthly GM = ______

4. Health ratios

Metric	Value	Target	Red flag
LTV / CAC	______	≥ 3×	< 2×
CAC payback period (months)	______	< 18	> 24
Gross margin %	______	≥ 70%	< 60%
Net revenue retention (12m forward)	______	≥ 120%	< 100%

5. 12-month scenario (fill after M3)

Re-forecast every month.

Month	New logos	Churn	MRR	Cumulative revenue	Burn	Cash position
M3 (today)	___	0	___	___	___	___
M4	___	___	___	___	___	___
M5	___	___	___	___	___	___
M6	___	___	___	___	___	___
M7	___	___	___	___	___	___
M8	___	___	___	___	___	___
M9	___	___	___	___	___	___
M10	___	___	___	___	___	___
M11	___	___	___	___	___	___
M12	___	___	___	___	___	___

6. Red-flag diagnostics

If a ratio is off, diagnose in this order (cheapest fix first):

Symptom	First suspect	Diagnostic
GM < 60%	LLM spend	Review model_routing dashboard; downgrade non-reasoning calls
GM < 60%	Infra waste	k6 baseline says p95 vs actual load — overprovisioned?
CAC too high	Founder-only sales	Is anyone else closing? if not, motion is not yet repeatable
CAC too high	Lead mix	Cold outbound vs warm referrals ratio — shift if skewed
LTV/CAC < 3	Contract length	Are pilots renewing verbally but not signing annual? Why?
Payback > 24	Pricing	Van Westendorp says prices could be higher — test
NRR < 100	Churn	Exit interview every churn — real reason, not stated reason

7. Evidence pointer

Do not fill in this worksheet from memory. Sources of truth:

Revenue: Stripe / invoice ledger
LLM spend: backend/app/services/model_routing_dashboard.py monthly export
Infra: AWS cost explorer export
CS/Support hours: weekly time logs
CAC: founder calendar + marketing ledger

Cite the source file/export in the margin when filling each row.

3.5 KiB Raw Blame History Unescape Escape