Pilot 1 Commitments
Pilot 1 is the first 50 paying clients, run over 6 weeks, ending in a single decision gate with three measured pass conditions. The gate determines whether we ramp to Pilot 2 (200-500 clients) or hold the cohort, keep collecting revenue, and course-correct. This sheet is the load-bearing reference document for that gate. The NewCo / JV memo points back to these exact criteria.
Scope, Duration, Gate
| Cohort | First 50 paying BOQ clients |
| Duration | 6 weeks from kickoff date |
| Gate | Single review at week 6, three pass conditions, all required |
| Pass condition 1 Conversion |
>= 1% of bot conversations close an action (booking, lead capture, payment, qualified handoff). Niche-specific definition agreed in writing before kickoff. |
| Pass condition 2 Churn |
<= 8% across the 50-client cohort over 6 weeks |
| Pass condition 3 COGS |
Modal client all-in <= $45/mo, heavy <= $63/mo, no client breaches the per-workspace daily cost cap |
| Pass all three | Ramp to Pilot 2 (200-500 clients) |
| Miss any one | No ramp. Cohort keeps earning. Tony and Donal course-correct on the failing condition before re-gate. |
You can verify each condition independently: conversion from the workspace dashboard, churn from Stripe, COGS from the credit-burn ledger.
What Ships in Pilot 1
| Deliverable | Status entering Pilot 1 | Pilot 1 commitment |
|---|---|---|
| 7 production-grade niche ICP packs | not built | All 7 authored and live before client #1 |
| Niche-pack authoring kit | designed | Donal can author packs without engineering touch |
| Workspace dashboard UI (per wireframe) | designed | Built to spec: three-column, KPI strip, edit-mode, role permissions |
| Per-workspace daily token budget + cost cap | designed | Live before client #100. Hard breaker at 2x cap. |
| CMS connector: Static HTML | scaffolded | Production-hardened |
| CMS connector: Shopify | OAuth landed | Connector live (Composio + auth-config) |
| CMS connector: Webflow | OAuth landed | Connector live (Composio + auth-config) |
| CMS connector: WordPress | not built | Native REST connector live |
| Customer-facing audit timeline UI | not built | Live, 90-day retention configured |
| 7-ICP classifier, 0.85 confidence threshold | scaffolded | Threshold raised, sub-threshold routed to Philippines fallback queue |
| Element-edit pipeline wired to live CMSes | demo only | Wired to the 4 Pilot 1 connectors above |
| Stripe production-hardening at 50+ concurrent renewals | scaffolded | Hardened: failure modes, idempotent webhooks, dispute handling |
Deferred to Pilot 2
These are explicitly out of scope for Pilot 1. Naming them now so nothing is implied:
- Written SLA (needs sustained-load telemetry from Pilot 1, then Pilot 2 to calibrate; defensible floor in Pilot 1 is 99.5% uptime, target post-Pilot-2 is 99.9%)
- CMS connectors: Wix, GoHighLevel, WooCommerce
- Squarespace (long tail, static-HTML fallback only)
- D1 sharding (single-DB carries Pilot 1; sharding is designed, not implemented)
- Voice channel
Substrate Brad Inherits on Day 1
Already landed. Not Pilot 1 work, not Pilot 2 work. Live in production today and forkable:
- Composio OAuth integration (landed 2026-05-12, rubric 0.79)
- Stripe scaffold supporting every BOQ tier shape ($1k / $1.5k / $2.5k / $700 upsells / $4k Brad retainer / one-time audits)
- Cascade schema for white-label (migration 0008_workspace_hierarchy)
- Credit-based metered billing ledger with 6 dedicated cron workers (migrations 0016-0018, plus auto-topup + allocation crons)
- Multi-bot Telegram channels (per-name token routing in claw/src/channels.ts)
- Pheromone substrate (claw/src/substrate.ts) - bot measurably improves on each client's business over time
- Branding API per workspace (logo, palette, fonts, dark/light, cascades down)
- github.com/one-ie/one MIT-licensed substrate (forkable by construction, bus-factor protection)
Mutual Commitments
BOQ (Brad)
- Surface first 50 client names + niche mix before kickoff
- Agree the niche-specific conversion definition in writing pre-kickoff
- Pay the $4k retainer per the contract
- Co-sign the week-6 gate decision
Online Optimisers (Donal)
- Run client-facing relationship + niche-pack QA
- Review and pre-edit the 7 niche-pack opening prompts
- Author additional niche packs using the authoring kit
- Co-sign the week-6 gate decision
ONE (Tony)
- Ship the technical deliverables in the table above
- Deliver weekly status against the gate metrics
- Source-code escrow for the product layer (NCC Group or equivalent)
- Co-sign the week-6 gate decision
What Happens at the Gate
Week 6 review. One meeting. Three numbers on the table.
Full pass
All 3 conditions met
Ramp to Pilot 2 (200-500 clients). Written SLA work begins. Wix / GHL / WooCommerce connectors enter build.
Partial pass (2 of 3)
One condition missed
No ramp. 2-week course-correction sprint on the failing metric. Re-gate at week 8. Cohort keeps earning.
Hard miss (2+ conditions missed)
Two or more missed
No ramp. Full pause. Tony + Donal + Brad sit down. Decide: re-scope Pilot 1, extend, or restructure.
In every outcome, the 50-client cohort keeps running and keeps earning. No client is shut off based on the gate. The gate decides what happens next, not what happens to the cohort.
Sign-off
This sheet is the source of truth for the Pilot 1 contract checkpoint. Changes require all three signatures.