Problem Define the customer problem, the vision, the business case and the measurement plan before anyone talks about a solution.
The foundational framework for problem definition. Be specific. Avoid solutions. Who has the problem, what it is, why it matters now.
RAPID for this section
Product VP RAI Product AVP RAI Product Leader RAID Product Manager RI Product Designer RI Content Designer RI BI Lead RI Tech Lead RI Agile Coach RI Research Lead RAPI Legal RI Compliance RI Finance RI
Agentic harness
Drafts Author Agent → FRAME Author · Prompt 1 · Draft ATask No solution language. WHO, WHAT and WHY describe the problem only. Attacks Adversarial Reviewer · Problem attack: is this a customer problem or an internal preference? Counterfactual: what happens if we do nothing for two quarters?Human check Confirm the problem is the customer's, not an internal preference.
The north-star statement: the customer outcome the product exists to change, and the success-criteria window.
RAPID for this section
Product VP RAI Product AVP RAI Product Leader RAID Product Manager RPI Product Designer RI Content Designer RI BI Lead RI Tech Lead RI Agile Coach RI Research Lead RAPI Legal RI Compliance RI Finance RI
Agentic harness
Drafts Author Agent → FRAME Author · Prompt 1 · Draft ATask Name the customer outcome, not the feature set. Attacks Adversarial Reviewer · Measurement attack: could the vision be 'achieved' while the customer is worse off?Human check Vision names an outcome a customer would recognize.
ROI estimate, Year-1 impact and risk profile, built by the Product Manager with the BI Lead. Must survive a downside case.
RAPID for this section
Product VP RAI Product AVP RAI Product Leader RAID Product Manager RPI Product Designer RI Content Designer RI BI Lead RI Tech Lead RI Agile Coach RI Research Lead RI Legal RI Compliance RI Finance RI
Agentic harness
Drafts Author Agent → FRAME Author · Prompt 1 · Draft ATask ROI and Year-1 arithmetic one line each; unsourced inputs = [GAP · BI Lead]. Attacks Adversarial Reviewer · Math attack: half the adoption, 1.5× the effort. Evidence attack: which input is asserted, not measured?Human check Every number traces to a source; the downside case is written down.
KPIs defined up front and tracked through ASSESS: metric, baseline (or 'not instrumented'), target, source. Required to exit G1.
RAPID for this section
Product VP RAI Product AVP RAI Product Leader RAID Product Manager RPI Product Designer RI Content Designer RI BI Lead RPI Tech Lead RI Agile Coach RI Research Lead RI Legal RI Compliance RI Finance RI
Agentic harness
Drafts Author Agent → FRAME Author · Prompt 1 · Draft ATask KPI = metric · baseline (or 'not instrumented') · target · source. Attacks Adversarial Reviewer · Measurement attack: does each KPI measure the stated problem, not a neighbouring one? Name the alternative cause.Human check BI Lead confirms baselines exist or are scheduled for instrumentation.
1.5 Stakeholder Map (RAPID) Name decision rights before Discovery starts. Default 13-role RAPID; name the individuals for this initiative. Product Leader is the sole decider.
RAPID for this section
Product VP RAI Product AVP RAI Product Leader RAID Product Manager RPI Product Designer RI Content Designer RI BI Lead RI Tech Lead RI Agile Coach RI Research Lead RI Legal RI Compliance RI Finance RI
Agentic harness
Drafts Author Agent → FRAME Author · Prompt 1 · Draft ATask Fill only 'Name for This Project'; exception approver = Product Leader. Attacks Adversarial Reviewer · Organizational attack: who can veto late and has not been told?Human check Named individuals agree they hold the letter next to their name.
1.6 DVQE — Desirability, Viability, Quality & Executability Score all four dimensions as evidence · source · verdict Strong / Partial / Weak. Required to exit G1; re-scored at 2.7.
RAPID for this section
Product VP RAI Product AVP RAI Product Leader RAID Product Manager RPI Product Designer RI Content Designer RI BI Lead RI Tech Lead PI Agile Coach RI Research Lead RI Legal RI Compliance RI Finance RI
Agentic harness
Drafts Author Agent → FRAME Author · Prompt 1 · Draft ATask Never leave a dimension unscored; a Weak verdict names the missing evidence as a [GAP]. Attacks Adversarial Reviewer · Evidence attack: grade each dimension measured / observed / self-reported / assumed / asserted.Human check Weak verdicts are named in the gate readout, not hidden.
G1 Stage gate · Problem Approved Confirms the problem is real, framed, sized — and worth solving.
RAPID at this gate
Product VP RAI Product AVP RAI Product Leader RAID Product Manager RPI Product Designer RI Content Designer RI BI Lead RI Tech Lead RI Agile Coach RI Research Lead RAPI Legal RI Compliance RI Finance RAI
Decides Product Leader Agree Product VP + Product AVP + Finance + Research Lead Agentic harness
Prepares Gate Reviewer Agent · Prompt 3 · Review. Checklist audit (Met / Partial / Missing), DVQE restated, key question answered, killer questions run, readout written in Appendix C2.Decides Product Leader · Advance to DISCOVER · Conditional · Hold / kill Gate checklists on the Process page
Research → PRFAQ → Customer Archetype Research the customer, write the press release and FAQ as if launching today, and re-score the evidence.
Write as if launching today. If the press release is not compelling, the product is not worth building.
RAPID for this section
Product VP RI Product AVP RAI Product Leader RAID Product Manager RI Product Designer RI Content Designer RI BI Lead RI Tech Lead RI Agile Coach RI Research Lead RAI Legal RI Compliance RI Finance --
Agentic harness
Drafts Author Agent → DISCOVER Author · Prompt 1 · Draft ATask Body = problem, solution, one role-attributed internal quote. Attacks Adversarial Reviewer · Problem attack: changed customer outcome or a feature list? Which benefit claim has no evidence?Human check Read aloud to one person outside the team; add their questions to 2.3.
2.2 Customer Quote & Getting Started A composite customer quote and the three-step getting-started story.
RAPID for this section
Product VP RAI Product AVP RAI Product Leader RAID Product Manager RI Product Designer RI Content Designer RI BI Lead RI Tech Lead RI Agile Coach RI Research Lead RAPI Legal RI Compliance RI Finance --
Agentic harness
Drafts Author Agent → DISCOVER Author · Prompt 1 · Draft ATask Quote is marked '(composite)'; never a real person. Attacks Adversarial Reviewer · Evidence attack: is a single anecdote doing the work of a dataset?Human check Getting started reflects what a first-time customer would actually do.
5–7 plain-language questions a customer would ask: pricing, access, eligibility, support, privacy.
RAPID for this section
Product VP RAI Product AVP RAI Product Leader RAID Product Manager RPI Product Designer RI Content Designer RI BI Lead RI Tech Lead RI Agile Coach RI Research Lead RAI Legal RI Compliance RI Finance --
Agentic harness
Drafts Author Agent → DISCOVER Author · Prompt 1 · Draft ATask 5–7 real customer questions, plain language. Attacks Adversarial Reviewer · Second-order attack: where is the escalation path when the product fails the customer?Human check Questions came from customers or outside readers, not the team.
2.4 Internal FAQ — Required Fields Standardized leadership questions answered the same way on every initiative. AI Risk Tier confirmed here.
RAPID for this section
Product VP RAI Product AVP RAI Product Leader RAID Product Manager RPI Product Designer RI Content Designer RI BI Lead RI Tech Lead RI Agile Coach RI Research Lead RAI Legal RI Compliance RI Finance --
Agentic harness
Drafts Author Agent → DISCOVER Author · Prompt 1 · Draft ATask AI Risk Tier = Project Identity value; one option per required field. Attacks Adversarial Reviewer · Organizational attack: is the AI Risk Tier understated? Who confirmed it?Human check Tech Lead confirms the AI Risk Tier; if Assistive or above, 3.5 is scheduled.
1–3 archetypes per initiative, derived from research. Archetypes, never real individuals. Not required to advance.
RAPID for this section
Product VP -- Product AVP -- Product Leader RAID Product Manager RI Product Designer RI Content Designer RI BI Lead RI Tech Lead RI Agile Coach RI Research Lead RAPI Legal RI Compliance RI Finance --
Agentic harness
Drafts Author Agent → DISCOVER Author · Prompt 1 · Draft ATask Archetypes, never a real person. Attacks Adversarial Reviewer · Evidence attack: interviews or a workshop? Would the competitor's PM say we misunderstood the customer?Human check Each archetype traces to at least five interviews, or the [GAP] is stated.
Front-stage (see / do / feel) and back-stage (systems, teams, handoffs) across a default six-stage model. Not required to advance.
RAPID for this section
Product VP -- Product AVP -- Product Leader RAID Product Manager RI Product Designer RI Content Designer RI BI Lead RI Tech Lead RI Agile Coach RI Research Lead RAPI Legal RI Compliance RI Finance --
Agentic harness
Drafts Author Agent → DISCOVER Author · Prompt 1 · Draft ATask The table's row names, including KPIs to Measure, are the fields. Attacks Adversarial Reviewer · Second-order attack: which back-stage handoff breaks first?Human check Journey stages match how the customer experiences the product.
2.7 DVQE — Desirability, Viability, Quality & Executability Re-score all four dimensions on DISCOVER evidence; explain any change from 1.6.
RAPID for this section
Product VP RI Product AVP RI Product Leader RAID Product Manager RI Product Designer RI Content Designer RI BI Lead RPI Tech Lead RPI Agile Coach RI Research Lead RAPI Legal RI Compliance RI Finance --
Agentic harness
Drafts Author Agent → DISCOVER Author · Prompt 1 · Draft ATask Re-score on DISCOVER evidence; explain changes from 1.6. Attacks Adversarial Reviewer · Evidence attack: did any verdict change from 1.6 without new evidence?Human check Changes from 1.6 are explained with evidence.
G2 Stage gate · PRFAQ Approved Confirms research, customer archetype, and launch story add up to a product worth designing.
RAPID at this gate
Product VP RI Product AVP RI Product Leader RAID Product Manager RPI Product Designer RI Content Designer RI BI Lead RI Tech Lead RI Agile Coach RI Research Lead RAPI Legal RAI Compliance RAI Finance --
Decides Product Leader Agree Legal + Compliance + Research Lead Agentic harness
Prepares Gate Reviewer Agent · Prompt 3 · Review. Checklist audit (Met / Partial / Missing), DVQE restated, key question answered, killer questions run, readout written in Appendix C2.Decides Product Leader · Advance to DESIGN · Conditional · Hold / kill Gate checklists on the Process page
Solution Intent and Validation → Capability Map Turn insights into a solution intent, a capability map and explicit boundaries. Evaluate AI risk when the product uses AI.
3–5 statements: 'As a [customer archetype], I learned that [observation], because [root cause].' Grounded in research, with implication and evidence tier.
RAPID for this section
Product VP -- Product AVP -- Product Leader RAID Product Manager RI Product Designer RI Content Designer RI BI Lead RI Tech Lead RI Agile Coach RI Research Lead RAPI Legal RI Compliance RI Finance --
Agentic harness
Drafts Author Agent → DESIGN Author · Prompt 1 · Draft ATask Format 'As [customer archetype], I learned that … because …' + Implication + Evidence tier. Attacks Adversarial Reviewer · Problem attack: which insight is a restated feature request?Human check No insight is a restated feature request.
3.2 Solution Intent and Validation Three-part conceptual solution in plain language, no tech specs: opportunity, why it works, riskiest assumption and its test.
RAPID for this section
Product VP -- Product AVP -- Product Leader RAID Product Manager RPI Product Designer RI Content Designer RI BI Lead RI Tech Lead RPI Agile Coach RI Research Lead RAPI Legal RI Compliance RI Finance --
Agentic harness
Drafts Author Agent → DESIGN Author · Prompt 1 · Draft ATask Opportunity = users, ROI range, fit; trade-off and riskiest-assumption test in Why It Works. Attacks Adversarial Reviewer · Counterfactual: name the cheaper non-product fix that captures most of the value. Does the first increment test the riskiest assumption?Human check Solution still answers the 1.1 problem; the riskiest assumption has a test.
Business capabilities the solution must enable, no implementation details. Each becomes one or more Features. Linked to insights (INS-IDs).
RAPID for this section
Product VP -- Product AVP -- Product Leader RAID Product Manager RPI Product Designer RI Content Designer RI BI Lead RI Tech Lead RPI Agile Coach RI Research Lead RAPI Legal RI Compliance RI Finance --
Agentic harness
Drafts Author Agent → DESIGN Author · Prompt 1 · Draft ATask Business capabilities, not features; A2 Reuse is the reuse field. Attacks Adversarial Reviewer · Problem attack: which capability is a screen in disguise? Which has no linked insight?Human check Every capability links to an insight; Tech Lead confirms feasibility in writing.
3.4 Boundaries & Constraints Won't-have list, hard constraints with owners, and what is explicitly out of scope.
RAPID for this section
Product VP -- Product AVP -- Product Leader RAID Product Manager RPI Product Designer RPI Content Designer RPI BI Lead RI Tech Lead RPI Agile Coach RI Research Lead RPI Legal RI Compliance RI Finance --
Agentic harness
Drafts Author Agent → DESIGN Author · Prompt 1 · Draft ATask Won't-have list is non-empty; each hard constraint has an owner. Attacks Adversarial Reviewer · Organizational attack: which hard constraint has no owner? Is the won't-have list empty?Human check Product Leader has reviewed the won't-have list.
3.5 AI Evaluation (Conditional) Required when the AI Risk Tier is Assistive, Decisional or Autonomous: AI role, failure modes, drift, evaluation and human oversight.
RAPID for this section
Product VP -- Product AVP -- Product Leader RAID Product Manager RPI Product Designer RPI Content Designer RPI BI Lead RI Tech Lead RPI Agile Coach RI Research Lead RPI Legal RI Compliance RI Finance --
Agentic harness
Drafts Author Agent → DESIGN Author · Prompt 1 · Draft ATask Project Identity tier governs; all four blocks if Assistive or higher, one line if None. Attacks Adversarial Reviewer · Second-order attack: what does the customer see on a confident wrong AI answer?Human check Compliance input is in the sources when the tier requires it.
G3 Stage gate · Build Ready Confirms the solution is approved, scoped, and prioritized. Engineering can pick it up.
RAPID at this gate
Product VP -- Product AVP -- Product Leader RAID Product Manager RPI Product Designer RI Content Designer RI BI Lead RI Tech Lead RAI Agile Coach RAI Research Lead RAPI Legal RAI* Compliance RAI* Finance --
Decides Product Leader Agree Research Lead + Tech Lead + Agile Coach (+ Legal + Compliance when AI Risk Tier ≥ Assistive) Agentic harness
Prepares Gate Reviewer Agent · Prompt 3 · Review. Checklist audit (Met / Partial / Missing), DVQE restated, key question answered, killer questions run, readout written in Appendix C2.Decides Product Leader · Advance to READY · Conditional · Hold / kill Gate checklists on the Process page
Roadmap → Sizing → Story Breakdown Build the roadmap, size the work, break it into Features and Stories, and hand it to the delivery tools.
Classify each capability Now, Next, Later or Parking lot, with the reason and target increment.
RAPID for this section
Product VP -- Product AVP -- Product Leader RAID Product Manager RPI Product Designer RI Content Designer RI BI Lead RI Tech Lead RI Agile Coach RPI Research Lead RI Legal RI Compliance RI Finance --
Agentic harness
Drafts Author Agent → READY Author · Prompt 2 · Draft BTask Classify each CAP into Now, Next, Later or Parking lot with reason and target increment. Attacks Adversarial Reviewer · Organizational attack: what falls off the roadmap to make room, and is that written down?Human check What falls off the roadmap to make room is written down.
The roadmap by increment with dependencies named to the providing team.
RAPID for this section
Product VP -- Product AVP -- Product Leader RAID Product Manager RPI Product Designer RI Content Designer RI BI Lead RI Tech Lead RI Agile Coach RPI Research Lead RI Legal RI Compliance RI Finance --
Agentic harness
Drafts Author Agent → READY Author · Prompt 2 · Draft BTask Fill the roadmap table from 4.1; dependencies name the providing team. Attacks Adversarial Reviewer · Organizational attack: which dependency owner has not agreed in writing?Human check Dependency owners have agreed in writing.
Initiative → Capability → Feature → Story → Task. Appendix A tables A1–A6 hold the handoff record.
RAPID for this section
Product VP -- Product AVP -- Product Leader RAID Product Manager RPI Product Designer RI Content Designer RI BI Lead RI Tech Lead RPI Agile Coach RPI Research Lead RI Legal RI Compliance RI Finance --
Agentic harness
Drafts Author Agent → READY Author · Prompt 2 · Draft BTask Treat A1–A6 as the tables to fill, one edit per table. Attacks Adversarial Reviewer · Evidence attack: which Feature has no falsifiable benefit hypothesis? Which Story fails INVEST?Human check Every Feature has a falsifiable benefit hypothesis and acceptance criteria.
4.4 Backlog Hygiene Standards Top two sprints refined; every Feature traces to a Capability and the Initiative; Stories pass INVEST; audit table of failures with owner and fix.
RAPID for this section
Product VP -- Product AVP -- Product Leader RAID Product Manager RPI Product Designer RI Content Designer RI BI Lead RI Tech Lead RI Agile Coach RPI Research Lead RI Legal RI Compliance RI Finance --
Agentic harness
Drafts Author Agent → READY Author · Prompt 2 · Draft BTask Record each failing Appendix A item: ID · standard failed · fix · owner. Attacks Adversarial Reviewer · Math attack: where is the missing 40% of effort (integration, compliance, testing, enablement, instrumentation)?Human check Zero open failures, or each has an owner and a date.
Where the brief ends and the delivery tools begin: Features, Stories and NFR tasks transferred, IDs pasted back, Definition of Ready checked.
RAPID for this section
Product VP -- Product AVP -- Product Leader RAID Product Manager RPI Product Designer RI Content Designer RI BI Lead RI Tech Lead RPI Agile Coach RPI Research Lead RI Legal RI Compliance RI Finance --
Agentic harness
Drafts Author Agent → READY Author · Prompt 2 · Draft BTask Leave the Ready to transfer verdict to the Reviewer (Prompt 3). Attacks Adversarial Reviewer · Measurement attack: which Story has no instrumentation event? Ready to transfer or Not ready, as a finding.Human check A Feature without an ID in A3 is not handed off.
G4 Stage gate · Product Approved Confirms the roadmap, sizing, and story breakdown are approved — architecture and design have signed off, and engineering can commit stories to a sprint.
RAPID at this gate
Product VP -- Product AVP -- Product Leader RAID Product Manager RPI Product Designer RI Content Designer RI BI Lead RI Tech Lead RAI Agile Coach RAI Research Lead RAI Legal RI Compliance RI Finance --
Decides Product Leader Agree Tech Lead + Agile Coach + Research Lead Agentic harness
Prepares Gate Reviewer Agent · Prompt 3 · Review. Checklist audit (Met / Partial / Missing), DVQE restated, key question answered, killer questions run, readout written in Appendix C2.Decides Product Leader · Advance to LAUNCH · Conditional advance (rework backlog) · Hold for resizing Gate checklists on the Process page
Agile → Test → Go/No-Go → Deploy Clear legal and compliance, deliver in sprints, test including accessibility, decide Go/No-Go, deploy with rollback ready, enable day one.
5.0 Legal & Compliance Review Legal and Compliance review before Go/No-Go, not after. 3.5 and 5.2 feed this packet. Never 'Cleared' without a source.
RAPID for this section
Product VP -- Product AVP -- Product Leader RAID Product Manager RPI Product Designer RI Content Designer RI BI Lead RI Tech Lead RI Agile Coach RPI Research Lead RAPI Legal RPI Compliance RPI Finance RI
Agentic harness
Drafts Author Agent → LAUNCH Author · Prompt 2 · Draft BTask Never write Cleared or Met without a source. Attacks Adversarial Reviewer · Evidence attack: 'Cleared' with no reviewer name and date is unsupported.Human check Every element Cleared or Cleared with conditions, with reviewer names and dates.
5.1 Agile Delivery (2-week sprints) Two-week sprints with planning, standup, review and retro. Scope locked at sprint start; backlog from READY feeds planning.
RAPID for this section
Product VP -- Product AVP -- Product Leader RAID Product Manager RPI Product Designer RI Content Designer RI BI Lead RI Tech Lead RPI Agile Coach RPI Research Lead RI Legal RI Compliance RI Finance RI
Agentic harness
Drafts Author Agent → LAUNCH Author · Prompt 2 · Draft BTask Fill the sprint table from Appendix A; stages not reached = [GAP: stage not reached · owner]. Attacks Adversarial Reviewer · Math attack: is capacity tracked or asserted?Human check Capacity and scope are tracked, not asserted.
5.2 Test Plan + Accessibility Functional, integration, performance, security and accessibility (WCAG AA). Accessibility cleared before Go/No-Go.
RAPID for this section
Product VP -- Product AVP -- Product Leader RAID Product Manager RI Product Designer RI Content Designer RI BI Lead RI Tech Lead RPI Agile Coach RPI Research Lead RI Legal RI Compliance RI Finance RI
Agentic harness
Drafts Author Agent → LAUNCH Author · Prompt 2 · Draft BTask Record coverage targets and evidence; accessibility evidence is attached, not promised. Attacks Adversarial Reviewer · Evidence attack: accessibility evidence attached or promised?Human check Accessibility evidence attached, not promised.
All P0 bugs resolved, test pass rate at target, accessibility cleared, observability instrumented, rollback rehearsed, 5.0 sign-off attached. Product Leader decides.
RAPID for this section
Product VP -- Product AVP -- Product Leader RAID Product Manager RPI Product Designer RI Content Designer RI BI Lead RI Tech Lead RPI Agile Coach RPI Research Lead RI Legal RI Compliance RI Finance RI
Agentic harness
Drafts Author Agent → LAUNCH Author · Prompt 2 · Draft BTask Each criterion Met / Not met / Unknown with the artifact that proves it. Attacks Adversarial Reviewer · Measurement attack: can every Go criterion fail? Any Unknown is a No-Go.Human check Every Go criterion Met with an artifact; no Unknown.
Phased deploy (canary → 10% → 50% → 100%) with rollback triggers tied to KPIs and error budgets. On-call for the first 48 hours.
RAPID for this section
Product VP -- Product AVP -- Product Leader RAID Product Manager RI Product Designer RI Content Designer RI BI Lead RI Tech Lead RPI Agile Coach RPI Research Lead RI Legal RI Compliance RI Finance RI
Agentic harness
Drafts Author Agent → LAUNCH Author · Prompt 2 · Draft BTask Rollback triggers are numbers tied to 1.4 KPIs or error budgets. Attacks Adversarial Reviewer · Second-order attack: a rollback trigger that never fired; triggers must be numbers tied to 1.4.Human check Rollback rehearsed with a date.
Runbooks, FAQs, training and comms ready; Day-1 dashboard live with the 1.4 KPIs; first retro within 30 days.
RAPID for this section
Product VP -- Product AVP -- Product Leader RAID Product Manager RPI Product Designer RI Content Designer RPI BI Lead RI Tech Lead RI Agile Coach RPI Research Lead RI Legal RI Compliance RI Finance RI
Agentic harness
Drafts Author Agent → LAUNCH Author · Prompt 2 · Draft BTask Fill the enablement table; dashboard KPIs are the 1.4 KPIs. Attacks Adversarial Reviewer · Second-order attack: a support spike with no runbook; a dashboard nobody watched.Human check Day-1 dashboard live and showing baselines; first retro on the calendar.
G5 Stage gate · Ship Ready Confirms the product is built, tested, accessible, legally cleared, and meets acceptance criteria and Definition of Done — ready to ship to production.
RAPID at this gate
Product VP -- Product AVP -- Product Leader RAID Product Manager RPI Product Designer RI Content Designer RI BI Lead RI Tech Lead RAI Agile Coach RAI Research Lead RI Legal RAI Compliance RAI Finance --
Decides Product Leader Agree Tech Lead + Agile Coach + Legal + Compliance Agentic harness
Prepares Gate Reviewer Agent · Prompt 3 · Review. Checklist audit (Met / Partial / Missing), DVQE restated, key question answered, killer questions run, readout written in Appendix C2.Decides Product Leader · Advance to production · Conditional · Hold Gate checklists on the Process page
Measure → KPIs → Scale / Fix / Sunset Read the KPIs set in FRAME, watch for drift, evaluate at 90 days and decide: Scale, Fix or Sunset.
6.1 KPI Tracking — activate the measurement plan Pull the 1.4 KPIs forward with baseline, 30/60/90-day actuals, variance and counter-metric. Owner reviews monthly.
RAPID for this section
Product VP RAI Product AVP RAI Product Leader RAID Product Manager RPI Product Designer RI Content Designer RI BI Lead RPI Tech Lead RI Agile Coach RPI Research Lead RI Legal RI Compliance RI Finance RI
Agentic harness
Drafts Author Agent → ASSESS Author · Prompt 2 · Draft BTask Treat Current as holding 30/60/90-day actuals, variance and counter-metric. Attacks Adversarial Reviewer · Measurement attack: did the metric move while the customer outcome did not? Counter-metrics present?Human check 90-day actuals for every 1.4 KPI, including counter-metrics.
6.2 Drift, Feedback Loop & Data Flywheel Track drift, capture corrections, retrain or revise on cadence. Signals → labels → improvements → retest, for AI and non-AI products.
RAPID for this section
Product VP RAI Product AVP RAI Product Leader RAID Product Manager RI Product Designer RI Content Designer RI BI Lead RPI Tech Lead RPI Agile Coach RPI Research Lead RAPI Legal RI Compliance RI Finance RI
Agentic harness
Drafts Author Agent → ASSESS Author · Prompt 2 · Draft BTask Treat Your Input and the harness signal table as the fields. Attacks Adversarial Reviewer · Second-order attack: model or behaviour drift with no owner and no cadence.Human check Drift signals have owners and a review cadence.
6.3 Post-Launch Evaluation (90-Day) Ninety days after launch: did the product meet its intent, yes or no, with evidence. Compares Year-1 impact to the 1.3 estimate.
RAPID for this section
Product VP RAI Product AVP RAI Product Leader RAID Product Manager RPI Product Designer RI Content Designer RI BI Lead RPI Tech Lead RI Agile Coach RPI Research Lead RPI Legal RI Compliance RI Finance RI
Agentic harness
Drafts Author Agent → ASSESS Author · Prompt 2 · Draft BTask Compare actuals to 1.3 and 1.4 explicitly; verdict yes / no with evidence. Attacks Adversarial Reviewer · Math attack: Year-1 actuals against the 1.3 estimate, explicitly.Human check 6.3 compares Year-1 impact to the 1.3 estimate explicitly.
Every product exits ASSESS through one of three doors. The Product Leader logs the decision, the number that drives it and the condition that would reverse it.
RAPID for this section
Product VP RAI Product AVP RAI Product Leader RAID Product Manager RPI Product Designer RI Content Designer RI BI Lead RPI Tech Lead RI Agile Coach RPI Research Lead RI Legal RI Compliance RI Finance RI
Agentic harness
Drafts Author Agent → ASSESS Author · Prompt 2 · Draft BTask Write 'Recommendation: …' only; owner and date = [GAP · Product Leader decides]. Attacks Adversarial Reviewer · Kill memo and steelman: is Fix really Sunset with extra steps? Is Scale justified by the KPI or by sunk cost?Human check Product Leader names one decision, the driving number and the reversal condition.
G6 Stage gate · Scale or Sunset Confirms the product met its intent at 90 days and exits ASSESS through one of three doors.
RAPID at this gate
Product VP RAI Product AVP RAI Product Leader RAID Product Manager RPI Product Designer RI Content Designer RI BI Lead RAI Tech Lead RI Agile Coach RI Research Lead RI Legal RI Compliance RI Finance RAI
Decides Product Leader Agree Product VP + Product AVP + BI Lead + Finance Agentic harness
Prepares Gate Reviewer Agent · Prompt 3 · Review. Checklist audit (Met / Partial / Missing), DVQE restated, key question answered, killer questions run, readout written in Appendix C2.Decides Product Leader · Scale · Fix and re-measure · Sunset Gate checklists on the Process page 