CUSTOMER EVIDENCE TO ENGINEERING BRIEF
Light Reach | October 12, 2026 | Editable template and illustrative example
Guide: https://www.lightreach.io/guides/customer-conversation-to-engineering-brief/
Prepared with AI assistance. Use authorized references for private evidence.

BLANK BRIEF

1. PROBLEM
Observed user failure:
Who is affected / how often, if known:
Why it matters:

2. EVIDENCE
Conversation or source reference:
Reproduction / observed behavior:
Relevant repository and revision:
Open questions and contradictory evidence:

3. OUTCOME
Behavior the user should experience after the change:
Observable completion rule:

4. OWNERSHIP
Product and repository:
Responsible owner:
Approval owner:
Why this is our team's work:

5. SCOPE AND CONSTRAINTS
Behavior to change:
Behavior and boundaries to preserve:
Shared rule / invariant and its owning boundary:
Analogous paths to inspect:
Non-goals:

6. ACCEPTANCE
Successful behavior:
Invalid input / failure behavior:
Repeated action / retry behavior:
Required environment and checks:
Missing dependencies or evidence:

7. DELIVERY
Reviewable diff:
Checks run and results:
Unresolved limitations:
Live deployment evidence, if release is in scope:

8. APPROVAL
Who may authorize starting work:
Who may approve the diff / merge:
Who may approve release:
What still needs a decision:

ILLUSTRATIVE EXAMPLE: INTERRUPTED IMPORT
This is a made-up scenario, not a real customer incident or measured outcome.

Problem: after an interrupted import, the customer does not know whether retry
will keep the original workspace and committed rows.
Source request: "Retry the import. Keep the original workspace."
Known context: a workspace identity and saved import checkpoint already exist.
Unknowns: exact interruption cause, checkpoint validity, and other affected paths.

Outcome: retry resumes from a valid checkpoint in the same workspace, preserving
committed rows. Repeating retry does not create duplicate rows. An invalid
checkpoint produces a clear recovery state instead of guessing a new workspace.

Investigation: locate the boundary that owns workspace identity and import
recovery. Inspect analogous retry paths before choosing a local patch.
Shared rule: preserve identity and committed rows across retry.
Scope: bounded recovery behavior. A broader onboarding redesign is separate work.

Acceptance evidence:
- Start with committed rows, interrupt, and retry; identity and rows remain.
- Repeat the retry; no duplicate rows are created.
- Use an invalid checkpoint; a clear recovery state is surfaced.
- Verify the relevant analogous recovery paths conform to the same rule.

Delivery: a diff and the actual checks run, with missing evidence recorded.
Approval: the responsible owner approves the concrete scope and release.
Outcome limits: passing tests proves tested behavior. Deployment and customer
friction are separate observations; no time saved or revenue gain is claimed.