> For the complete documentation index, see [llms.txt](https://knowledgebase.autorabit.com/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://knowledgebase.autorabit.com/product-guides/integral-zone/iz-suite/workspaces/code-analysis/ai-focus-report.md).

# AI Focus Report

The AI focus report is a tab in a repository branch's dossier. It shows an AI review of one assessment of a branch or pull request: what to fix first, a fix plan, the guidance the review followed, the evidence it relied on, and the decisions your team records afterwards. Use it to decide what to act on without reading every finding, and to plan the fix.

Open it from **Workspaces → Code Analysis → Branches & baselines**: open a repository, choose the branch or pull request in **Branch / pull request**, then select the **AI focus report** tab. When the selected assessment already has a completed AI review, the dossier opens on this tab.

{% hint style="info" %}
Requires the **Code Analysis** licence, AI turned on for your workspace, and a role that can view AI reviews. Requesting a review also needs a role that can run AI reviews.
{% endhint %}

<figure><img src="https://content.gitbook.com/content/9vAxMuDrkUkB4OXlH9CL/blobs/cdn1MB9QZALf5I7HbV03/code-ai-main.png" alt="AI focus report of a repository branch"><figcaption><p>AI focus report of a repository branch</p></figcaption></figure>

## How a Review Is Created

A review covers one assessment of one branch or pull request. It is created in one of two ways.

### Automatically After an Assessment

New assessments are reviewed automatically when all three of these are true:

1. AI is turned on for your workspace.
2. AI is allowed for the repository by your AI access rules.
3. An enabled AI review profile applies to the repository's asset type, at the global, platform or asset level.

### On Request

1. On the **AI focus report** tab, select **Request an AI review**. If the branch already has a report, select **Prepare fresh review**.
2. Check the **Review profiles applied**. Profiles apply from global, to platform, to asset level; where their guidance conflicts, the more specific level wins.
3. Optionally, add **Additional review input**: what you want the review to check or know, such as an intended change, its business impact or a constraint.
4. Select **Prepare evidence briefing**. IZ Suite gathers the native findings, metrics, repository details and, when source storage is on, the stored source code. The briefing stays within IZ Suite.
5. Review the inputs and the budget on the next screen, then approve sending the evidence to your configured AI provider and model.
6. The review runs. Its progress is shown on the tab, and the report appears when it completes.

<figure><img src="https://content.gitbook.com/content/9vAxMuDrkUkB4OXlH9CL/blobs/S6dHWKezedzMU7GBrqDx/code-ai-request.png" alt="Request an AI review drawer"><figcaption><p>Request an AI review</p></figcaption></figure>

{% hint style="warning" %}
Nothing is sent to the AI provider until you approve it, and each review is charged to your AI budget. See [Usage & Budgets](https://knowledgebase.autorabit.com/product-guides/integral-zone/iz-suite/workspaces/ai-and-automation/usage-and-budgets.md).
{% endhint %}

{% hint style="danger" %}
Do not enter passwords, tokens or other secrets in **Additional review input**.
{% endhint %}

### When AI Has Not Reviewed an Assessment

If the selected assessment has no review, the tab shows **AI has not reviewed this assessment** with a checklist of what automatic reviews need. Each step that is not met links to the screen where you fix it. When an earlier assessment of the branch was reviewed, the tab also links to that review and its fix plan.

<figure><img src="https://content.gitbook.com/content/9vAxMuDrkUkB4OXlH9CL/blobs/KriZiWpam6gSPUd2J5Dx/code-ai-notreviewed.png" alt="An assessment that AI has not reviewed"><figcaption><p>An assessment that AI has not reviewed</p></figcaption></figure>

## Reading the Report

The strip at the top shows the review status, how many evidence records the review used and how many evidence gaps it found. **View agent execution** shows how the review ran, and **Prepare fresh review** starts a new one.

The report has five tabs:

| Tab                      | What it shows                                                                                      |
| ------------------------ | -------------------------------------------------------------------------------------------------- |
| **Recommendations**      | What needs your attention, in priority order, with each recommendation and how to confirm the fix. |
| **Project fix plan**     | Related recommendations combined into one reviewed change per file, ready to apply together.       |
| **Review guidance**      | The review profiles and guidance this review followed.                                             |
| **Supporting evidence**  | What AI checked and the evidence the review is based on.                                           |
| **Decisions & activity** | Decisions and follow-up recorded by your team, earlier reviews, and usage.                         |

### Recommendations

**What needs your attention** summarises the review: the number of priorities and the AI model that produced it. **Open native findings** opens the assessment's own findings, and **Export review** downloads the review.

Above the list, a banner shows how many Blocker and Critical findings remain open. When the review has questions for you, **Answer & re-review** lets you answer them and run the review again with that context.

<figure><img src="https://content.gitbook.com/content/9vAxMuDrkUkB4OXlH9CL/blobs/H6GnJN0ax2LtRA2FgQOp/code-ai-attention.png" alt="What needs your attention summary, open findings banner and filters"><figcaption><p>What needs your attention</p></figcaption></figure>

Filter the list with **Needs attention** or **All recommendations**, and by AI priority. Select a recommendation to see:

| Section                           | Description                                                                                                                                                         |
| --------------------------------- | ------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
| **Native findings**               | The assessment findings the recommendation relates to, each linked to its finding. A recommendation with no matching finding is marked as an additional AI concern. |
| **What AI found**                 | The problem, and why it matters.                                                                                                                                    |
| **Recommended fix**               | The steps to fix it.                                                                                                                                                |
| **Review coordinated fix**        | The proposed change for each affected file.                                                                                                                         |
| **How to confirm the fix**        | How to check that the fix worked, usually a rescan and a test.                                                                                                      |
| **What still needs verification** | What the review could not confirm from the evidence.                                                                                                                |
| **Affected configuration / code** | The lines the recommendation refers to, from the assessment's snapshot.                                                                                             |

<figure><img src="https://content.gitbook.com/content/9vAxMuDrkUkB4OXlH9CL/blobs/OBxDMkXKoNmLKPasq90p/code-ai-recommendation.png" alt="Recommendation list with the selected recommendation&#x27;s details"><figcaption><p>A recommendation</p></figcaption></figure>

### Project Fix Plan

The fix plan starts with a short summary of where to begin. It then combines the recommendations into one change per file and counts the proposed file changes, the files inspected and the recommendations with work remaining. **Download combined patch** downloads all proposed changes together, and **Review coverage and limits** lists what the plan does and does not cover.

### Review Guidance

Shows the review profile and revision used, and each guideline the review applied, from the broadest level (global) to the most specific (asset). The report keeps the guidance as it was when the review ran; changes you make to a profile apply to future reviews. To change the guidance, edit the [AI Review Profiles](https://knowledgebase.autorabit.com/product-guides/integral-zone/iz-suite/operations/quality-control/ai-review-profiles.md).

<figure><img src="https://content.gitbook.com/content/9vAxMuDrkUkB4OXlH9CL/blobs/r7LlZP9QONU0s6sXiGSY/code-ai-guidance.png" alt="Guidance used for this review"><figcaption><p>Review guidance</p></figcaption></figure>

### Supporting Evidence

Lists **What AI checked**, the questions that still need your input, and the **Captured evidence** the review used, such as the repository context, inventory metrics, the selected assessment, quality gates and findings. Search, refresh and export the list, and open any record.

<figure><img src="https://content.gitbook.com/content/9vAxMuDrkUkB4OXlH9CL/blobs/wAcdasiKidibqePZqyS0/code-ai-evidence.png" alt="Supporting evidence with the captured evidence list"><figcaption><p>Supporting evidence</p></figcaption></figure>

### Decisions & Activity

Records the decisions and follow-up your team adds to the review, and links to earlier reviews of the same assessment. From here you can view the review's usage and audit trail, and **Remember a verified lesson** for future reviews of the repository.

<figure><img src="https://content.gitbook.com/content/9vAxMuDrkUkB4OXlH9CL/blobs/Q0a2kX0fSBgNjWJqqu0z/code-ai-decisions.png" alt="Human decisions and follow-up"><figcaption><p>Decisions and activity</p></figcaption></figure>

## Things to Know

* **The assessment does not change.** AI recommendations and proposed changes do not close findings or change the quality gate. Fix the findings, or record an approved exception, then rescan.
* **Proposed changes are proposals.** Review each change, then build, test and rescan the branch before you rely on it.
* **A missing record is not proof.** If the evidence does not show a control, dependency or vulnerability, that does not mean it is absent. Check the evidence gaps and the coverage limits.
* **A review belongs to one assessment.** After a new scan of the branch, the earlier review stays linked to its own assessment. Request a fresh review to cover the new one.

{% hint style="info" %}
When AI access allows AI for a repository but not its source code, the review uses only its findings, metrics and context, and no source code is sent.
{% endhint %}

## Next Steps

* [Branches & Baselines](https://knowledgebase.autorabit.com/product-guides/integral-zone/iz-suite/workspaces/code-analysis/branches-and-baselines.md): open a repository's dossier.
* [AI Review Profiles](https://knowledgebase.autorabit.com/product-guides/integral-zone/iz-suite/operations/quality-control/ai-review-profiles.md): set the guidance reviews follow and which repositories are reviewed automatically.
* [AI Access](https://knowledgebase.autorabit.com/product-guides/integral-zone/iz-suite/workspaces/ai-and-automation/ai-access.md): choose which repositories AI may review, and whether it may see their source code.
* [Usage & Budgets](https://knowledgebase.autorabit.com/product-guides/integral-zone/iz-suite/workspaces/ai-and-automation/usage-and-budgets.md): set AI budgets and see what reviews cost.
* [AI Coverage & Findings](https://knowledgebase.autorabit.com/product-guides/integral-zone/iz-suite/workspaces/ai-and-automation/ai-coverage-and-findings.md): see AI review coverage across all applications.


---

# 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://knowledgebase.autorabit.com/product-guides/integral-zone/iz-suite/workspaces/code-analysis/ai-focus-report.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.
