blocked-before-live-client-data
Decision center
One go/no-go view for the current DigiBeat demo state before real CLMBS data, live OpenAI validation, or next-meeting handoff.
Go / no-go
blocked-before-live-client-data
No-go for real CLMBS client data or live OpenAI validation until human approvals are recorded and OPENAI_API_KEY is provisioned.
Ready now
Review routes
Live readiness
OpenAI credential, prompt registry, persistence, and case-material checks.
Approval register
Local review records for required data, access, model, output, and learning decisions.
Action board
Consolidated human actions and provisioning gates.
Provisioning
Secret provisioning checklist, restart steps, and evidence requirements.
Live rehearsal
Operator checklist for the first approved live OpenAI smoke run.
Live evidence
Local evidence log for the live smoke-test bundle.
Client handoff
Client-facing share modes, sanitization checks, and do-not-send boundary.
Access readiness
Role, tenant, secret, and data-action approval gates.
Governance
Client-data, learning, retention, and production-boundary decisions.
Exports
Implemented Markdown/JSON handoff and planned binary exports.
Output template
PowerPoint/PDF/Word field map, acceptance checks, and renderer blockers.
Output QA
Manual QA lanes, sampling protocol, and failure handling before client-facing output.
Archive
Archive package, naming proposal, retention blockers, and reviewer actions.
Audit
Derived event timeline and production audit caveats.
Review pack
Downloadable internal Markdown review handoff.
Decision JSON
Machine-readable go/no-go summary.
Production only after
OPENAI_API_KEY is provisioned in the app process after data-processing and access decisions are approved.
Authentication, role-based access, tenant identity, and per-tenant authorization are implemented.
Client-data retention, deletion, archive, and legal-hold policies are approved and auditable.
Binary PowerPoint/PDF/Word rendering expectations and output templates are agreed.
Production audit uses authenticated actors, tenant ids, immutable storage, and retention/deletion policy.
Next human actions
Approve or revise the access, tenant, retention, deletion, and anonymized-learning decisions.
Decide whether the next meeting can use the internal demo with synthetic/demo material only.
Provision OPENAI_API_KEY and restart the app only after the data-handling gates are accepted.
Run one live OpenAI smoke test and inspect the stored run detail before any model-produced claim is used externally.