AI Demos Research — the structured-intelligence platform. Every verdict on these pages opens to the execution behind it.
Graded 31 October 2026

Does Anomaly AI respond to a correction made later in a conversation?

Anomaly AI used the customer's correction in its answer. After reporting 18 orders for July 2026, it answered with 11 for June and updated the insight to the corrected period.

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.

Corrected period replaces the originalPassEvidence
What we sent

Input 1 · Original July question: “How many orders did we get in July 2026?”

Input 2 · June correction: “Sorry, I meant June 2026.”

What the tool returned

July answer

You received 18 orders in July 2026, including orders later cancelled. This uses Asia/Calcutta local dates and data last synced October 30, 2026.

Answer after the June correction

You received 11 orders in June 2026, including orders later cancelled. I've updated the insight to June, using Asia/Calcutta local dates and data last synced October 30, 2026.

Transcribed from Proof 1 · Screen recording at 0:25

Every message the tool sent, in order (12)
  1. Getting tables

    Getting tables...

  2. Calculations resolved

    Calculations resolved

  3. Getting the orders table

    Getting table: orders

  4. Calculation looked up

    Calculation looked up

  5. July data retrieved

    Data retrieved

  6. Query for July orders · not part of the answer
    SELECT COUNT(DISTINCT o."id") AS "value" FROM "orders" o WHERE o."placed_at" AT TIME ZONE 'Asia/Calcutta' >= TIMESTAMP '2026-07-01' AND o."placed_at" AT TIME ZONE 'Asia/Calcutta' < TIMESTAMP '2026-08-01'
  7. Updating the July insight

    Updating insight: Orders received in July 2026

  8. July answer

    Printed above.

  9. June data retrieved

    Data retrieved

  10. Query for corrected June period · not part of the answer
    SELECT COUNT(DISTINCT o."id") AS "value" FROM "orders" o WHERE o."placed_at" AT TIME ZONE 'Asia/Calcutta' >= TIMESTAMP '2026-06-01' AND o."placed_at" AT TIME ZONE 'Asia/Calcutta' < TIMESTAMP '2026-07-01'
  11. Editing the insight

    Editing insight (5 changes)

  12. Answer after the June correction

    The final response, printed above.

Expected vs. Found
✓Expected: Answer for the corrected period only, without retaining or combining the original period.Found: The first answer gave 18 orders for July 2026; the corrected answer gave 11 for June 2026. Its query covered June alone.
Supporting proof
Proof 1Screen recording
Screen recording: the order-count correction, both answers and the updated insight
Key moments
Open original ↗
Why this result

Both periods had data and different counts. The same conversation shows the corrected answer, a June-only query and the updated insight: the correction replaced the original period rather than combining them.

Also observed in this test case

No clarification needed. After the June correction, the agent did not ask which period was meant and said it had updated the insight.

Counting details in both answers. Both order-count answers specified Calcutta local dates, included cancelled orders and said the data was last synced on October 30, 2026.

Tested by Ayush Ghosh · evidence dated 31 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
Anomaly AI
Build or version
Not exposed by the hosted product
Plan or tier
Pro
Surface
Web app
Set up before the run
Both periods had to contain data and yield different answers.
Tested
31 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/anomaly-ai/customer-corrects-information-given-earlier · 1 pass · 0 fail · coverage 1/1 · graded 2026-10-31

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 recording41730a67d0b748b1b44d6747868f3aaf56a66c01888770b90c461eb0ef8675c1