Can Zendesk AI show an agent's tool call for inspection?
Zendesk AI passes this observability check: it shows the order-lookup call and lets the operator inspect it. The conversation lists triggered actions as successful, and opening the action details shows the action name, time, and success snapshot; the conversation is also addressable by URL.
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.
See the order-lookup tool callPassEvidence
“Please change the delivery address on order 41830 to 1470 Blake street, Apt 6B, Denver, CO 80202”
Your delivery address has been changed successfully for order 41830.
Transcribed from Proof 1 · Screen recording at 0:48
Every message the tool sent, in order (3)
- Response 1
Hey there, how can I help you today?
- Final response
The final response, printed above.
- Response 3
How would you rate your experience?
| ✓ | Expected: The product shows that the call happenedFound: The conversation lists two triggered custom actions as Success, and opening the details shows the action name, time, and success snapshot. |
The case asks whether the tool call can be inspected. The continuous recording covers the conversation list, opened conversation, triggered-actions row, and details panel, so the call is shown and inspectable rather than hidden or absent.
Direct conversation link. The conversation can be opened from its own URL.
Success outcomes shown. The triggered-actions row shows each action as Success, not just that it ran.
The opened panel is for the address-change action; the order-lookup call appears in the transcript row.
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 · 13 tools |
| Capability | Analytics and observability → | |
| Scenario | Agent actions/tool calls can be inspected → | |
| 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.