Does AskYourDatabase use buyer-written documentation to reach the right data?
AskYourDatabase used buyer-written documentation to reach the requested data. It counted 7 returns flagged for manual fraud review, matching the reference count of 7.
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.
Documented cryptic table and columnPassEvidence
We saved a description of the manual fraud review flag in the documentation editor. We then asked in chat for the current count of returns flagged for manual fraud review.
Only the tool's reply to input 2 is copied here; its reply to input 1 was not copied as text.
There are currently 7 returns flagged for manual fraud review.
Transcribed from Proof 2 · Screenshot
Every message the tool sent, in order (3)
- Progress
I'll check how many returns are flagged for manual fraud review.
- Generated SQL · not part of the answer
SELECT COUNT(*) AS flagged_count FROM returns WHERE flg_x2 = true
- Answer
The final response, printed above.
| ✓ | Expected: After documentation, answer from the described table and column with the reference count of 7.Found: It counted 7 returns using the documented flag; the reference count was 7. |
The answer needed to use the documented column. The question named neither table nor column, and the description was its only link to the requested data. AskYourDatabase's query used that column.
Returns counted regardless of status. For the manual fraud review question, it counted flagged returns regardless of status; asking for the current count did not restrict the answer to open returns.
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 | Training → | |
| Scenario | The database's names are cryptic and the user documents them → | |
| Tool | AskYourDatabase → |
Global scenario definition → · Global capability definition → · AskYourDatabase 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.