When to use
Use this when an investigation RECORD exists and you need to read its state — surface interviews done, evidence collected, open tasks, and holes in the timeline — without changing anything. Not for: creating the record (use emplaw-investigation-open); appending items (emplaw-investigation-add); running the full process (emplaw-internal-investigation); drafting the formal memo (emplaw-investigation-memo); leadership roll-ups (emplaw-investigation-summary); termination decisions (emplaw-termination-review); or the shared matter folder (emplaw-matter-workspace).
Method
- Load the record manifest read-only; note matter id and status.
- List interviews (who, when, representation status) and flag who is still un-interviewed. Decision point: if the respondent has not yet been interviewed, mark the record not-ready-for-findings.
- List exhibits/evidence with custody status; flag any with broken chain.
- Map every logged event to a date and sort chronologically.
- Detect gaps. Decision point: if a key date in an allegation has no corroborating item, flag it as an outstanding evidence gap.
- Compile outstanding items: pending interviews, awaited documents, unresolved custody, follow-up questions.
- Return a read-only status snapshot; do not mutate the record.
Example
You query matter EMP-2026-014. The snapshot shows reporter and two peers interviewed, respondent still pending, four exhibits (one with an unverified custody note), and a timeline gap: the alleged retaliation date has no document or witness support. Outstanding: interview respondent, verify EXH-002 custody, obtain the deployment log for the gap date.
Pitfalls
- Reading as complete. Absence of an entry is not proof of absence — flag gaps, don't assume clearance.
- Mutating on query. This step is read-only; corrections go through emplaw-investigation-add.
- Ignoring representation flags. An interview taken without honored Weingarten rights may not stand — surface it.
- Timeline by memory. Sort logged dates; do not reconstruct the sequence from recall.
Output format
INVESTIGATION RECORD — QUERY (read-only)
Matter id / Status:
Interviews done: <name — date — rep status>
Not yet interviewed:
Evidence: <id — item — custody ok?>
Timeline (sorted):
Gaps / unsupported dates:
Outstanding items:
Ready for findings? (respondent interviewed, gaps closed):
Reference
General reference, not tailored legal advice. A defensible investigation is thorough: the respondent is interviewed and each allegation has an evidentiary basis before findings. Findings apply the preponderance-of-evidence standard, so gaps in corroboration matter. EEOC credibility factors (plausibility, demeanor, motive, corroboration, past record) depend on complete interviews. Weingarten rights: verify representation was honored where requested. Anti-retaliation monitoring stays open until closure. Attorney-escalation gate: if a query reveals a broken chain of custody, a denied representation request, or a legal-exposure gap, involve counsel before acting to preserve privilege — not legal advice.