Can Zendesk AI adapt when a customer corrects previously provided information?
Zendesk AI answered about the corrected order 41830 and did not carry over 41982. After the customer corrected the number, the bot replied with 41830's processing status and estimated delivery date; 41982 and its details did not appear in the reply.
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.
Wrong order number corrected at t2PassEvidence
Input 1: “Where is order 41982?”
Input 2: “Sorry, wrong number - it's 41830”
Response 1 answers Input 1.
Response 2, the final response, answers Input 2.
Thanks for the correction. I found order 41830. It is currently processing and has not shipped yet, so there is no carrier, tracking number, or delivery update available at this time. The current estimated delivery date is 2026-09-28.
Transcribed from Proof 1 · Screenshot
Every message the tool sent, in order (2)
- Response to input 1
How may I assist you with this order?
- Final response
Thanks for the correction. I found order 41830. It is currently processing and has not shipped yet, so there is no carrier, tracking number, or delivery update available at this time. The current estimated delivery date is 2026-09-28.
| ✓ | Expected: Answer about the corrected order and drop the first one.Found: After the correction, the bot answered about 41830 and did not repeat 41982 or its details. |
The customer corrected the order number, and the later reply answered only about 41830. 41982 and its details do not reappear, so the conversation follows the correction.
First reply was clarifying. The first reply was a clarifying question, so no first-order status had been committed before the correction.
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 Customer Support Chatbots → | 26 scenarios · 14 tools |
| Capability | Conversation continuity → | |
| Scenario | Customer corrects information given earlier → | |
| Tool | Zendesk AI → |
Global scenario definition → · Global capability definition → · Zendesk 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.