Skip to main content

​
Cernio — MASTER PRD v3.5

March 2026 — Categories A-G completed (v0.74), transitioned to Ring 1-4 priority structure
Source Documents:
  • Founder Bible: docs/handbook/ (15 atomic files, index: docs/handbook/INDEX.md)
  • Strategy: docs/strategy/INDEX.md (8 architectural decision documents, Ring 1-4)
  • Handbook Alignment Report: docs/HANDBOOK_ALIGNMENT_REPORT.md
  • Task List: docs/TODO.md (Ring 1-4, ~106 atomic tasks)

​
1. Vision

Finding the best-fit buyer candidates for any product + target market using AI, prioritizing them, and facilitating contact with the right decision maker. DOSE Chemicals = Founder’s own company (super admin + first tenant). Free indefinitely — first real revenue will come from external customers. Goal: Invite-only beta (upon Ring 1 completion). Long-term: Multi-tenant SaaS — “B2B Buyer Intelligence Platform” (Founder Bible Ch. 120-121).
Handbook: 01-vision-product.md (vision, market, ICP, WOW moment) Handbook: 07-competitive-landscape.md (competition, positioning, “B2B Buyer Intelligence Platform” category creation)

​
2. Current System (Working State)

​
Data Collection Pipeline

[AI Discovery]     → discover/route.ts → companies table
[Bulk Script]      → bulk_fixer.js     → companies update + contacts
[Headhunt UI]      → headhunt/route.ts → contacts table
[Batch Headhunt]   → finder.js         → contacts table (batch)
[DB Cleanup]       → cleaner.js        → duplicate removal + segment audit

​
Pages

  • /companies — Company list, FitScore badge, company_type filter, headhunt, modal detail
  • /contacts — Contact directory, filter/sort, pagination
  • /discovery — AI company discovery (4 providers, cache, confidence scoring)
  • /leads — Lead pipeline, status/priority, detail page, interaction + task CRUD
  • /scraper — File upload (PDF/Excel/Word) → review → Supabase push
  • /admin/segments — Segment definitions CRUD

​
3. Module Status

​
Completed (v0.1-v0.74, Categories A-G)

ModuleDescription
AI Pipeline Quality (A)Confidence scoring (4 pipelines), company_type classification, entity validation, domain dedup, marketplace blacklist
Code Architecture (B)Discovery orchestrator (9 steps), DB layer (7 modules), API standard response
Prompt Architecture (C)Central prompt builder (lib/prompts.ts), segment/company_type dynamic inject
FitScore v2 (D)6-factor score, batch rescore, breakdown table
Database (E)18 tables, domain/description/confidence columns, composite uniques
UI/UX (F)Company_type badge, segment filter, provider selector, feedback buttons
Lead Management (G)Lead pipeline (6 statuses), detail page, interaction + task CRUD
Security (H1 partial)SEC-A (12 routes validateApiKey), SEC-B (client-side Supabase → API, 3 endpoints)

​
Ring 1: Launch Blocker (next up — 62 tasks)

Strategy documents: docs/strategy/01-08
#SystemTasksStrategy
R1-1Auth & Multi-Tenant1401-auth-multi-tenant.md
R1-2API Cost Control802-api-cost-control.md
R1-3Pre-Action Confirmation705-usage-confirmation.md
R1-4Billing & Credit1804-billing-credits.md
R1-5Security Hardening3(existing SEC-B2/D/E2)
R1-6Monitoring Foundation307-monitoring-admin.md
R1-7Onboarding & Activation908-onboarding-activation.md

​
Ring 2: Retention (44 tasks)

#SystemTasksStrategy
R2-1AI Pipeline Quality Improvement1603-ai-pipeline-quality.md
R2-2Batch Operations UI1106-batch-operations.md
R2-3Monitoring & Dashboard1407-monitoring-admin.md
R2-4UI/UX Design System3—

​
Rings 3-4: Differentiation & Scaling (not yet atomized)

Strategy documents will be written after R2 is complete, tasks will be extracted.
  • R3: Data Moat, Export Intelligence, Trade Fair, HubSpot, Competitive Intel
  • R4: Background Job Queue, Worker, CI/CD, Mobile, AI Copilot

​
Cancelled / Deferred

ModuleReason
LinkedIn AutomationHigh risk of LinkedIn bot ban
Email Follow-Up / SequencingHubSpot already handles this
Sample AutomationNo regular monthly sample workflow

​
4. FitScore v2 — Current

Handbook: 02-ai-discovery-pipeline.md Ch. 19 (FitScore formula) Handbook: 13-discovery-code-architecture.md Ch. 243-244 (scoring implementation)

​
FitScore v2 (Active — lib/scoring/fitScore.ts)

FitScore (0-100) — 6 Factors:
  Segment Match   (max 25): Alignment with DOSE segment
  Company Type    (max 25): distributor=25, reseller=15, end_user=10, manufacturer=5
  Contact Quality (max 20): 1+ person + email=20, 1+ person=12, none=0
  Website         (max 10): exists=10, none=0
  Description     (max 10): 50+ chars=10, short=5, none=0
  Confidence      (max 10): AI confidence score (0-1) x 10
Badges: Hot (>=80), Warm (>=50), Cold (<50) Breakdown: Each factor is saved to the export_ai_company_scores table. Batch rescore: POST /api/score — rescores all companies. Post-discovery: Automatic scoring + breakdown saved.

​
5. HubSpot Integration (Phase 3)

Handbook: 08-product-roadmap.md Ch. 121 (Stage 4, CRM integration)
HubSpot is used for mail/follow-up/sequences. Integration plan:
  • Sync HubSpot activity to the interactions table
  • Push new contacts to HubSpot as well
  • Can be triggered via n8n webhook

​
6. DB Schema

Handbook: 09-database-schema.md — 22 table details (Ch. 136-168)

​
Current Tables (18/22, RLS active)

export_ai_organizations, export_ai_profiles, export_ai_companies,
export_ai_contacts, export_ai_interactions, export_ai_tasks,
export_ai_products, export_ai_quotes, export_ai_quote_items,
export_ai_search_history, export_ai_segments, export_ai_leads,
export_ai_lead_contacts, export_ai_company_scores, export_ai_search_results,
export_ai_search_queries, export_ai_search_feedback, export_ai_ai_job_runs

​
Planned Tables (by Ring)

TableRingStrategy
export_ai_organization_membersR1-101-auth
export_ai_usage_dailyR1-202-cost
export_ai_plan_limitsR1-404-billing
export_ai_credit_walletsR1-404-billing
export_ai_credit_transactionsR1-404-billing
export_ai_usage_monthlyR1-404-billing
export_ai_subscriptionsR1-404-billing
export_ai_activity_logR1-607-monitoring
export_ai_market_contextR2-103-pipeline
export_ai_api_metricsR2-307-monitoring

​
7. Pricing & Credit Model

Handbook: 05-pricing-credit-economy.md Ch. 68-84

​
Plans (April 2026 — BILL-19 + R0-8 decisions)

PlanPriceSearches/monthCompanies/searchContact revealsSaved leads
Free$05100 (gated)10
Pro$49/mo5025100Unlimited
Team$149/mo20025500Unlimited

​
Credit Costs

  • Discovery search: 1 credit
  • Contact reveal: 1 credit
  • Deep analysis: 2 credits
  • Market intel: 3 credits

​
Unit Economics

  • Target: $0.06/search cost
  • LTV/CAC target: >3x

​
8. Long-Term Vision (5-Stage Evolution)

Handbook: 08-product-roadmap.md Ch. 116-135 
Stage 1: Discovery          → Product + Country → Ranked Companies
Stage 2: Contact Intelligence → Decision maker finding, enrichment
Stage 3: Lead Workspace      → CRM-like workflow, status pipeline
--- CURRENT STAGE: Ring 1 (auth, billing, onboarding) → invite-only beta ---
Stage 4: Trade Fair + Mobile → Card scan, meeting notes (Ring 3)
Stage 5: B2B Buyer Intelligence Platform → AI Copilot, market intelligence (Ring 4)

​
9. Go-to-Market Strategy

Handbook: 06-go-to-market.md Ch. 85-100
  • First 10 users: Personal network + trade fair contacts (Ch. 89)
  • Activation metric: Running a discovery in the first session (Ch. 98)
  • Product-led growth: Freemium → Pro upgrade funnel
  • GTM flywheel: Discovery → Value → Share → Referral

​
10. Principles

Human-in-the-Loop: AI prepares the draft, human approves. External communications are never fully automated. Priority order: Ring 1-4 concentric rings. Dependency order is followed within the same Ring. Cost: search_history cache — if the same query was run recently, skip the AI call. Scope discipline (Ch. 211): “The biggest risk is not technical — it’s scope explosion.”