CodexWorkBluefin customer portal: discovery brief
Designed example · synthetic data. This is an authored illustration, not a recorded agent run. All organizations, people and figures are fictional.
Purpose: explore a portal so customers can check request status without calling (sponsor S3). Authorized work: discovery only — “Looks good, go ahead on discovery.” (S1). Build and release remain unapproved (S4, C2).
Scope
Can proceed under the recorded discovery approval: organize the supplied notes, map known status-chasing questions and compare status-only and payments options. This is proposed discovery work, not a new sponsor commitment.
Needs another input or decision: inspecting CRM fields requires an authorized export; committing to phase-one scope requires the sponsor to resolve the Operations/Sales disagreement.
Constraint: keep phone intake available (O1). Not yet decided: online payments; Operations favors status viewing while Sales proposed payments (O2). Outside current approval: production build, spending and launch (C2).
Ownership and timing
Marta Quill is the named sponsor. No implementation owner is recorded (O4). IT is the source of the export dependency, not an assigned delivery owner (C1). “Before the holiday rush” is a requested timeframe; launch date not set (S2, C3). No build budget or extra headcount is approved (S4, O4).
Proposed sequence — no committed dates
| Stage | Output | Entry condition |
|---|---|---|
| 1. Start discovery from supplied material | Known questions, constraints and scope alternatives | Discovery is already approved; use the supplied notes |
| 2. Inspect CRM export | Field and data-quality notes | Authorized export is actually available |
| 3. Recommend phase-one scope | Status-only and payments tradeoffs | Separate known facts from export-dependent questions; sponsor chooses scope |
| 4. Decide build | Explicit go/no-go and resources | Separate approval for scope, owner, budget and target date |
Risks: missing export blocks field inspection, not all discovery; an undefined baseline prevents measuring fewer calls; premature launch language can be mistaken for build authority.
Next decisions: settle the scope before committing to a phase-one plan; identify the authorized export owner and date before field inspection. Build resources and launch timing are later decisions, required before build rather than before the already-approved discovery. Decision list · Commitment table.