Open this record at pilot entry for a customer or internal service. Handoff is complete only when the receiving team demonstrates the operating capability; documentation delivery alone is insufficient. When the building team retains service ownership, record that decision and have its designated backup demonstrate the capabilities without the original builder. Do not invent an organizational transfer or an exit requirement for a properly staffed, permanent service team.
Ownership
| Responsibility | Primary | Backup | Escalation | Evidence exercised |
|---|---|---|---|---|
| Business outcome and scope | — | — | — | — |
| Product/workflow roadmap | — | — | — | — |
| Service and on-call | — | — | — | — |
| Architecture and release | — | — | — | — |
| Agent harness and behavior configuration | — | — | — | — |
| Data and context quality | — | — | — | — |
| Tools and integrations | — | — | — | — |
| Identity, policy, and risk | — | — | — | — |
| Evaluation and model behavior | — | — | — | — |
| Adoption instrumentation and value measurement | — | — | — | — |
| User support and training | — | — | — | — |
Pilot transfer plan
| Capability | Delivery lead demonstrates | Paired operation | Operating team or backup leads | Acceptance evidence | Gap owner / due |
|---|---|---|---|---|---|
| Harness, model route, prompt, context policy, and guardrails | — | — | — | — | — |
| Tool, identity, authorization, egress, and effect controls | — | — | — | — | — |
| Evaluation cases, graders, fixtures, and reports | — | — | — | — | — |
| Adoption event, denominator query, guardrails, and rebaseline | — | — | — | — | — |
| Operator workflow, access, training, support, and adoption diagnosis | — | — | — | — | — |
| Release, canary, rollback, and dependency lifecycle | — | — | — | — | — |
| Alerts, incidents, reconciliation, and support | — | — | — | — | — |
Capability evidence
| Capability | Demonstration | Pass evidence | Date | Approver |
|---|---|---|---|---|
| Explain scope, limits, and architecture | Design review led by receiving team | Questions and risks resolved | — | — |
| Add a representative evaluation case | Case authored and run independently | Valid fixture, grader, and expected result | — | — |
| Change the harness safely | Version one behavior component and run affected-route tests | Compatible manifest, evaluation, canary, and rollback evidence | — | — |
| Reproduce adoption measurement | Run numerator and eligible-denominator queries from declared sources | Result matches dashboard and exclusions are explained | — | — |
| Enable one representative cohort | Receiving team grants access, runs workflow-native training, and supports first use | Users reach the final surface through the owned path; access and support evidence is retained | — | — |
| Diagnose an adoption break | Trace one weak funnel transition through representative cases and reason codes | Owning layer, response, due date, and follow-up measure are accepted | — | — |
| Release a compatible change | Branch, review, promotion, and canary | Release evidence and healthy soak | — | — |
| Roll back behavior | Trigger simulated rollback | Prior version restored and verified | — | — |
| Respond to an alert | Game day from detection to containment | Alert, owner, kill switch, and readback work | — | — |
| Reconcile an external effect | Source-of-truth comparison | Affected state identified and corrected | — | — |
| Manage access and policy | Joiner/mover/leaver and policy change | Least privilege and audit verified | — | — |
| Support an operator | Ticket triage and feedback loop | Resolution and backlog classification | — | — |
| Review value and adoption | Monthly service review led by owner | Decision and follow-up actions recorded | — | — |
| Retire the workflow | Tabletop decommission | Identity, tools, state, audit, and users covered | — | — |
Capability-transfer and dependency evidence
Measure whether the receiving team can operate and change the service without hidden delivery-team work. For a retained internal service, measure dependence on the original builder and unplanned support instead. Falling temporary delivery involvement is useful only while accepted outcomes, adoption, safety, reliability, and full cost remain healthy.
| Measure or exercise | Baseline | Exit target | Observed result and window | Evidence source | Owner / decision |
|---|---|---|---|---|---|
| Representative changes completed by the receiving team without delivery-team intervention | — | — | — | — | — |
| Routine support cases resolved without delivery-team intervention | — | — | — | — | — |
| Delivery-team hours per accepted outcome or compatible change | — | — | — | — | — |
| Unresolved target-specific or parallel production assets | — | 0 or approved time-bounded exceptions | — | — | — |
| Product/platform gaps with an accepted owner and disposition | — | — | — | — | — |
Document temporary support with a scope, owner, expiry, and exit exercise. Do not manufacture self-sufficiency by shifting unmeasured work or risk to operators. ADP-002, FDE-004, OPS-003.
Knowledge and assets
- [ ] Workflow charter, value case, current-state map, and decision log
- [ ] Architecture, domain model, tool and policy contracts, and data lineage
- [ ] Threat model, evaluation suites, reference worlds, and known limitations
- [ ] Release manifest, environments, dependency lifecycle, and rollback procedure
- [ ] Dashboards, alerts, SLOs, runbooks, kill switches, and reconciliation queries
- [ ] User guide, training material, support intake, and escalation routes
- [ ] Backlog classified as customer configuration, reusable pattern, platform gap, model limitation, operating problem, or retirement candidate
- [ ] Every field-built production asset has a target, product/platform, or retirement owner and normal change path
- [ ] Reuse rights and transfer constraints are recorded for customer-funded, jointly developed, or confidential work
Artifact ownership and lineage
| Artifact/release role | Authoritative URI | Version or digest | Upstream lineage | Current change owner | Receiving owner | Promotion path | Access/retention | Last verified |
|---|---|---|---|---|---|---|---|---|
| Workflow charter and value/adoption contract | — | — | — | — | — | — | — | — |
| Data, domain, and context contracts | — | — | — | — | — | — | — | — |
| Harness and behavior bundle | — | — | — | — | — | — | — | — |
| Tool, identity, and policy contracts | — | — | — | — | — | — | — | — |
| Evaluation suite and release evidence | — | — | — | — | — | — | — | — |
| User surface, telemetry, runbooks, and support assets | — | — | — | — | — | — | — | — |
Every production row resolves to an immutable version or digest and names both the upstream inputs and the path used to promote a change.
Post-exit support and re-entry contract
Complete this section when temporary delivery support ends. For a retained internal service team, use the owned service and succession contract instead; record why temporary-support fields are inapplicable.
| Field | Value |
|---|---|
| Receiving-team support hours, channel, severity, and response target | — |
| Temporary delivery-team support scope and maximum hours | — |
| Temporary support owner and expiry | — |
| Escalation and succession path | — |
| First post-exit service review | — |
| Re-entry trigger and approving owner | New workflow evidence / severe incident / material policy or dependency change / other |
| Explicitly excluded delivery-team work | Routine operation / ordinary support / unowned backlog / other |
Re-entry is a scoped decision, not a reset to permanent embedded delivery. Preserve the receiving team's ownership, name the new evidence or event, bound the work, and set the next exit exercise.
Acceptance decision
Record accepted, accepted with dated conditions, or not accepted. Include remaining gaps, temporary delivery support if any, final service owner, next service review, and rollback or retirement trigger.