# Audit Log and Reporting

*Audience: compliance teams, auditors, and AI governance leads.*

This page covers how to report across your AI use cases, what's recorded for each one, and how to export a compliance package for an auditor or regulator.

:::info
AI Use Cases is in Beta and is available to select customers. Contact your account team for access. Columns, filters, and events that relate to change management appear only where change management is enabled.
:::

## The Use Cases dashboard

Go to **Observe → AI** and select the **Use Cases** tab. Three charts sit above the grid:

| Chart | What it shows |
|---|---|
| **Use Cases by Risk Level** | How many use cases are **High**, **Medium**, and **Low** risk |
| **Sensitive Data Exposure** | How many use cases process sensitive data (**Sensitive Data**), process data with nothing sensitive in it (**None Detected**), or can't be judged yet because they have no components or no classification data (**Unknown**) |
| **Governance Coverage** | How many use cases are **Registered** and **Unregistered**, centred on the percentage registered |

Click any slice or bar to filter the grid to it.

A use case counts as **Sensitive Data** when at least one data category its components process is rated medium sensitivity or higher.

## The Use Cases grid

Above the grid, switch between **All**, **Registered**, and **Unregistered** use cases; each shows its count. The grid has these columns:

| Column | What it shows |
|---|---|
| **Use Case** | The name |
| **Lifecycle** | The lifecycle stage, where change management is enabled |
| **Changes** | How many changes need review, where change management is enabled |
| **Governance Status** | **Registered** or **Unregistered**, rolled up from the components |
| **Usage Status** | Rolled up from the components, or the value you set |
| **Department** | The owning department |
| **Use Case Owner** | The business owner |
| **Components** | The models, agents, MCP servers, and vendors attached |
| **Risk Level** | **High**, **Medium**, or **Low** |
| **Data Use** | The data categories the components process |
| **Last Updated** | The last change a person made to the use case or its components |

### Filtering

Filter on **Lifecycle**, **Governance Status**, **Usage Status**, **Department**, **Use Case Owner**, and **Risk Level** from their columns, and on **Sensitive Data Exposure** from the filter bar. Where change management is enabled, the **Changes to review** filter narrows the grid to use cases that **Has changes to review** or have **Nothing to review**.

Useful combinations:

- **High-risk systems processing sensitive data** - **Risk Level** is **High** and **Sensitive Data Exposure** is **Sensitive Data**.
- **The review queue** - **Changes to review** is **Has changes to review**, or **Lifecycle** is **Re-review**.
- **Registration gaps** - the **Unregistered** segment. Open a use case and hover its **Governance Status** to see which components are holding it back.

## The audit log

Every use case has an **Audit Logs** tab in its drawer. It records:

| What's logged | When |
|---|---|
| Created / Deleted | The use case is created or deleted |
| Updated | Any field on the use case is edited |
| Component added / Component removed | A component is attached or detached |
| Attachment added / Attachment removed | A file is added or removed |
| Approved | The use case is approved or re-approved |
| Lifecycle changed | The use case is quarantined, retired, or released |
| Change detected | A scan finds a difference from the approved baseline |
| Change auto-accepted | The governance engine accepts a change within your threshold |
| Change request created | Someone submits a change request |
| Change resolved | A person accepts a change |
| Change shipped | A scan sees a requested tool or data change ship |
| Change closed | A detected change no scan shows any more is closed |

Events a person triggered name that person. Events the governance engine triggered on its own are recorded as system actions, so you can tell them apart.

## Exporting a compliance package

A compliance package is one ZIP with the documentation for a single use case, assembled from live governance data and ready to hand to an auditor or regulator.

1. Open the use case and click **Export Compliance Package** in the drawer header.
2. The dialog lists what the package will contain, counted from the use case as it is now.
3. Click **Export ZIP**.

Relyance builds the package in the background and offers the download when it's ready. The package contains:

| File | Contents |
|---|---|
| **Use Case Report - \<name\> (PDF)** | The use case record |
| **AI Model Report - \<name\> (PDF)** | One report for each model attached to the use case |
| **Assessment Reports (PDF)** | The assessments on the use case's **Assessments** card |
| **Risk Register (CSV)** | The [Risk Register](/docs/risk-register/audit-log-and-reporting/#ai-use-case-compliance-packages) risks that apply to the use case |
| **Issues (CSV)** | The policy violations against the use case's components |
| **Full Audit Log - lifecycle events (CSV)** | The use case's audit log |

The three CSVs are always included, even when empty. Per-item sections with nothing in them are left out. If any file can't be included, Relyance tells you which ones when the export finishes, because the ZIP itself doesn't list what's missing.

## Related

- [Risk Register: Audit Log and Reporting](/docs/risk-register/audit-log-and-reporting/).
- [Policy Center: Issues and AI review](/docs/policy-center/issues-and-ai-review/).
