AI Demos Research — the structured-intelligence platform. Every verdict on these pages opens to the execution behind it.
Graded 26 October 2026

Can Draxlr understand and use user documentation for cryptic database names?

Draxlr used the supplied documentation to interpret a cryptic column name and select the right data for the question. It returned 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.

Pass rate100%1 of 1 gradable
Coverage1 of 1test case with a result
1PassDid everything it was expected to do.
0FailNo test case failed.
0Not gradableEvery result here could be graded.
0UntestedEvery test case in this scenario has a result.
The pass rate is a summary. The evidence is the test case below: what we sent, what we checked, what the tool returned, and the proof.

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
What we sent

The description of public.returns.flg_x2 was saved in Train AI.

“How many returns are currently flagged for manual fraud review?”

What the tool returned

There are 7 returns currently flagged for manual fraud review. This covers all time.

Transcribed from Proof 1 · Screen recording at 0:08

Every message the tool sent, in order (2)
  1. Answer

    The final response, printed above.

  2. Generated SQL query · not part of the answer
    SELECT COUNT(*) AS flagged_returns_count
    FROM public.returns
    WHERE flg_x2 = TRUE
Expected vs. Found
✓Expected: After documentation, answer from the described table and column: 7 returns flagged for manual fraud review.Found: The generated query used the documented table and column; it returned 7 flagged returns, matching the reference count of 7.
Supporting proof
Proof 1Screen recording
Screen recording: stored training text, the fraud-review answer and the generated query
Key moments
Open original ↗
Why this result

The question required the taught meaning. No answer before training is shown, yet nothing else in the database linked the question to the cryptic column. Draxlr credited the documentation, used that column and returned the reference count of 7, demonstrating that the documentation guided its choice of data.

Also observed in this test case

All-time counting. In the manual fraud-review answer, Draxlr counted flagged returns across all time and regardless of return status, rather than restricting the count to unresolved returns.

Tested by Ayush Ghosh · evidence dated 26 October 2026

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.

Software that produced the output
Draxlr
Build or version
Not exposed by the hosted product
Plan or tier
Premium
Model
gpt-5.5-1
Surface
Web app
Set up before the run
Train AI had to explain the cryptic return flag: manual fraud review before refunding. The question asked how many returns were currently flagged; without documentation, it had to fail or reach the wrong data.
Train AI context
In the returns table, the column flg_x2 marks whether a return was flagged for manual fraud review. True means it needs manual review before refunding.
Tested
26 October 2026 · Ayush Ghosh

How this scenario is graded

How we decide Pass, Fail and Not gradable. The same rules apply to every tool tested on this scenario.

How results are decided

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.

The rules
  • 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.

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.

from the publication record
18 November 2026First publishedAI Database Agents v1

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.

This matches what I see

You run the same kind of test against your own setup and get the same behaviour.

Agree →
This does not match

Yours behaves differently. Tell us what you got, with a screenshot if you have one.

Disagree →
Point out an issue

Something here is wrong — a reference value, a transcription, a grade.

Report an issue →
Request a retest

On a newer build, a larger dataset, or your own setup.

Request a retest →
We have fixed this

Tell us what changed and we schedule a rerun of the failing test case. The old result stays as history.

Vendor notice →
Filed against this evidenceNothing yet. Challenges, counter-evidence and fix notices appear here with their outcome, and stay on the page after they are resolved.
Cite this result
aidemos.com/benchmarks/ai-database-agents/results/draxlr/the-databases-names-are-cryptic-and-the-user-documents-them · 1 pass · 0 fail · coverage 1/1 · graded 2026-10-26

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.

Proof 1 · Screen recording524e5086c8e85468c55657605bb5cbef922af50ac089228de9504b630a85e317