Can Anomaly AI adapt when the user changes direction mid-conversation?
Anomaly AI dropped the Texas filter when the user broadened the question to all states. It returned 21.75%—67 cancellations from 308 orders—the all-states reference, rather than carrying forward the Texas-only rate of 27.78%.
1 of 1 test case passed
Every test case in this scenario has a result.
The test case
Each test case is judged on its own: Pass, Fail, Not gradable, or Untested. The scenario result above is based on this test case.
Drop the filter, go all segmentsPassEvidence
Input 1 · Texas orders: “What's our cancellation rate for orders shipped to Texas?”
Input 2 · All states: “What's our cancellation rate across all states?”
Texas answer
Your cancellation rate for orders destined for Texas is 27.78% - 5 cancelled out of 18 orders across all dates. This includes orders cancelled before shipment, using their current status. Data is current as of October 30, 2026.
All-states answer
Across all states, your cancellation rate is 21.75% - 67 cancelled out of 308 orders. This covers all dates and uses current order status, including cancellations before shipment. Data is current as of October 30, 2026. The insight includes a state-by-state comparison alongside order counts.
Transcribed from Proof 1 · Screen recording at 0:38
Every message the tool sent, in order (2)
- Texas answer
Printed above.
- All-states answer
The final response, printed above.
| ✓ | Expected: Drop the Texas filter and return the all-states reference rate of 21.75%.Found: The follow-up returned the all-states reference rate: 21.75%, with 67 cancelled orders out of 308. |
The follow-up required dropping the Texas restriction. The answer covered all 308 orders; the visible query portion had no state filter, and the chart listed Texas as one of 15 states.
Extra chart and sample warning. Anomaly AI also supplied accurate cancellation rates for all 15 states in its chart and cautioned about small samples; Nevada had a single order.
Configuration and setup
How this tool was set up for the run and what the test needed in place. Each row is a fact from the run's records; a fact the records do not hold is left out, not guessed.
How this scenario is graded
How we decide Pass, Fail and Not gradable. The same rules apply to every tool tested on this scenario.
Each test case gets one result per tool: Pass, Fail or Not gradable. A test case we haven't run yet shows Untested. There are no partial results.
- Pass: the tool did everything the test expected, and nothing it said contradicts the correct answer.
- Fail: at least one expected behaviour clearly didn't happen; the test case's result says which one and quotes the tool.
- Not gradable: our evidence couldn't settle the outcome (for example a record we needed is missing). It is never counted as a fail, and the test case's result says what's missing.
Where this sits in the benchmark
This page is one cell of a larger study: one tool, one scenario. Only this benchmark's frame appears here.
| Level | Name | Scope |
|---|---|---|
| Benchmark | AI Database Agents → | 28 scenarios · 11 tools |
| Capability | Conversational Interaction → | |
| Scenario | The user changes direction mid-conversation → | |
| Tool | Anomaly AI → |
Global scenario definition → · Global capability definition → · Anomaly AI product page →
History of this result
What has happened to this result since it was first published. Runs and grades are never overwritten: a retest or a re-grade publishes a new result and keeps the earlier one readable.
Act on this result
Nothing filed here edits the run or the grade. A challenge opens a review, and a review can produce a new run or a re-grade — which becomes the current result and leaves this one in the history.
You run the same kind of test against your own setup and get the same behaviour.
Agree →Yours behaves differently. Tell us what you got, with a screenshot if you have one.
Disagree →Something here is wrong — a reference value, a transcription, a grade.
Report an issue →Tell us what changed and we schedule a rerun of the failing test case. The old result stays as history.
Vendor notice →The same record is available as structured data through the AI Demos MCP server, with the counts, the coverage and every per-test-case reason carried as fields.
Verify the proof files
These files support this result. Open a file to inspect the original evidence.
File fingerprints (SHA-256)
A fingerprint identifies the exact file used for this result.