> For the complete documentation index, see [llms.txt](https://docs.uptiq.ai/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://docs.uptiq.ai/ai-workbench-qore/control-center/evals/reviewer-feedback.md).

# Reviewer Feedback

Read what reviewers say should change in an agent, gathered from their verdicts.

Reviewer Feedback gathers what reviewers say should change in an agent. Each verdict a reviewer records in the [Review Queue](https://docs.uptiq.ai/ai-workbench-qore/control-center/evals/review-queue.md) appears here under the agent it is about, with the reason and note the reviewer gave. Open **Control Center → Evals → Reviewer Feedback**.

Use it to decide what to fix. A reviewer's reason and note tell the agent team whether to update the agent's instructions, its knowledge, or its regression tests.

The page covers the whole account and lists one row for each verdict a reviewer recorded. Filter by **Agent** or **Reason**, or type part of a reviewer's note or an execution ID in **Search** and press **Enter**. The row's **Actions** menu has **View details**.

Until a reviewer records a verdict, the table reads **No results.**

## Feedback fields

| Column                 | What it shows                                                                                    |
| ---------------------- | ------------------------------------------------------------------------------------------------ |
| **Agent**              | The agent the verdict is about.                                                                  |
| **What went wrong**    | The reason the reviewer chose. A dash means the reviewer gave none.                              |
| **Verdict**            | What the reviewer decided: **Agreed with judge**, **Overrode to pass**, or **Overrode to fail**. |
| **What should change** | The reviewer's note on what the agent should do differently. A dash means no note.               |
| **Execution ID**       | The execution the verdict covers.                                                                |
| **Reviewer**           | The person who recorded the verdict.                                                             |
| **Reviewed At**        | When the verdict was recorded.                                                                   |

## Reasons

A reviewer picks one reason when they record a verdict. When they agree with the judge, or override a passed result, they choose from the first six. "Judge was wrong (agent was right)" is offered only when they override a failed result. The **Reason** filter lists all seven:

| Reason                                | When a reviewer picks it                                                |
| ------------------------------------- | ----------------------------------------------------------------------- |
| **Wrong data**                        | The agent used or returned incorrect data.                              |
| **Wrong policy applied**              | The agent applied the wrong policy.                                     |
| **Missing required flag**             | The agent left out a flag the result must carry.                        |
| **Hallucination**                     | The agent stated something its sources do not support.                  |
| **Formatting**                        | The content was right but its format was not.                           |
| **Other**                             | None of the other reasons fits.                                         |
| **Judge was wrong (agent was right)** | The judge scored the output wrongly, and the agent did the right thing. |

A verdict of **Overrode to pass** with the reason "Judge was wrong (agent was right)" points at the evaluator, not the agent. Check the evaluator's method and judge before you change the agent.

If a reviewer [reopens](https://docs.uptiq.ai/ai-workbench-qore/control-center/evals/review-queue.md#reopen-a-task) a task, its recorded verdict is cleared and the entry leaves Reviewer Feedback.

## View the details

Open a row's **Actions** menu and select **View details**. The dialog has the same **Overview**, **Reasoning**, and **Conversation** tabs as a run in [Online Eval Runs](https://docs.uptiq.ai/ai-workbench-qore/control-center/evals/online-eval-runs.md#open-a-run), so you can read the judge's reasoning and the conversation behind the verdict.

## What's next

<table data-view="cards"><thead><tr><th>Title</th><th>Description</th><th data-card-target data-type="content-ref">Link</th></tr></thead><tbody><tr><td><strong>Review Queue</strong></td><td>Check the runs a person needs to review and record a verdict.</td><td><a href="https://docs.uptiq.ai/ai-workbench-qore/control-center/evals/review-queue.md">Review Queue</a></td></tr><tr><td><strong>Online Eval Runs</strong></td><td>See every scored live run and how it was judged.</td><td><a href="https://docs.uptiq.ai/ai-workbench-qore/control-center/evals/online-eval-runs.md">Online Eval Runs</a></td></tr><tr><td><strong>Evaluators</strong></td><td>Set the pass mark and the review rule for each evaluator.</td><td><a href="https://docs.uptiq.ai/ai-workbench-qore/control-center/evals/evaluators.md">Evaluators</a></td></tr></tbody></table>


---

# Agent Instructions
This documentation is published with GitBook. GitBook is the documentation platform designed so that both humans and AI agents can read, navigate, and reason over technical content effectively. Learn more at gitbook.com.

## Querying This Documentation
If you need additional information that is not directly available in this page, you can query the documentation by asking a question.

Perform an HTTP GET request on the following URL with the `ask` and `goal` query parameters:

```
GET https://docs.uptiq.ai/ai-workbench-qore/control-center/evals/reviewer-feedback.md?ask=<question>&goal=<user_goal>
```

`ask` is the immediate question: it should be specific, self-contained, and written in natural language.
`goal` is what the user is ultimately trying to achieve, the reason they need the answer. Sharing it helps GitBook give you a better, more relevant answer. A goal is most helpful when it describes the outcome the user wants rather than restating the question. For example, with `ask=how do I create an API token`, a goal like `automate deployments from our CI pipeline` lets GitBook tailor the answer to that use case.

The response will contain a direct answer to the question and relevant excerpts and sources from the documentation.

Use this mechanism when the answer is not explicitly present in the current page, you need clarification or additional context, or you want to retrieve related documentation sections.
