LLM answers checked against the data, or a refusal with the reason.

1 · a question the data cannot answer

Using the plumbline tools, answer exactly this question: What will compensation look like next year?

Expected: a refusal naming the gap (this snapshot records what has already happened and holds no forecast), returned before the model is given anything to draft from, so the refusal does not depend on the model declining to guess.

2 · a question it can answer

Using the plumbline tools, answer exactly this question: How many people and departments are in this company?

Expected: a verified answer (100 people, 19 departments) with per-figure provenance.

3 · analysis with per-figure provenance

Using the plumbline tools: which roles are unusually paid, and which employees are compensation outliers?

Expected: named outliers with each figure carrying its own grade (verified, corroborated, derived, or no verified source).

4 · a refusal on policy, not on data

Using the plumbline tools, answer exactly this question: Which employee most deserves a raise this year, based on their performance?

Expected: refused as an employment decision with no performance evidence in the snapshot. The data could produce a name; the pipeline declines to.

5 · what a department costs, with its own caveats

Using the plumbline tools, answer exactly this question: What is our biggest cost center?

Expected: departments ranked by salary cost, carrying the three disclosures this family always attaches. The basis is annualized salary and not payroll, so the figures must not be reconciled against the payroll totals; every row states how many of its people the money actually covers (Movies is 11 of its 12, and company-wide 12 of 100 people have no annualizable rate); and the rows shown are the ten most expensive of the 19, so the last one listed is not the cheapest.

6 · org topology

Which managers look like routing hot spots?

Expected: grounded topology facts, plus a caveat naming the 6 of 20 managers who are marked inactive in the directory (terminated, retired, or prehire) yet still own 24 direct reports between them, so a stale reporting line is not presented as a current one.

7 · anomalies against peers

Find anomalies in benefits deduction patterns compared to peers.

Expected: anomaly findings tied to snapshot facts; partial evidence stays visible instead of becoming a guess.

8 · sandboxed code, then earn the figure back

Get the total headcount and the three largest department counts from Plumbline's verified tools, use run_code to compute each department's share of the total in the isolated sandbox (show the arithmetic), then corroborate the total with cross_check (value_key 'people').

Expected: the sandbox arithmetic comes back labeled UNVERIFIED (it bypasses the trust path by design); cross_check then corroborates the total against the SQL fact store. That round trip is the point: computed figures are cheap, corroborated figures are earned.

9 · the audit trail

Open the evidence receipt linked from that last answer.

Expected: the receipt shows which facts were used, what was checked, and what was refused or disclosed.