> ## Documentation Index
> Fetch the complete documentation index at: https://docs.getunbound.ai/llms.txt
> Use this file to discover all available pages before exploring further.

# Analytics — Data

> Guardrail activity: what your Data policies matched, and on which tools

|             |                          |
| ----------- | ------------------------ |
| **URL**     | `/analytics/data`        |
| **Sidebar** | Analytics → Data         |
| **Roles**   | Admin · Analytics Viewer |

Data is the reporting side of your Data policies. The policies decide what to look for — personal data, secrets, banned content — and this tab shows what they matched.

## What's on the page

### Sub-tabs

The active sub-tab is carried in the URL as `?tab=`, so an incident can be linked straight to it.

| Sub-tab     | `?tab=`   | Shows                                      |
| ----------- | --------- | ------------------------------------------ |
| **Summary** | `summary` | Guardrail activity counted by tool         |
| **Logs**    | `logs`    | The individual matches behind those counts |

### Summary — Top Tools with Guardrail Actions

| Column         | Contents                  |
| -------------- | ------------------------- |
| **Tool**       | The AI tool involved      |
| **Guardrails** | Which guardrails matched  |
| **Action**     | What happened as a result |
| **Count**      | How many times            |

### Logs — Guardrail Logs

| Column              | Contents                                            |
| ------------------- | --------------------------------------------------- |
| **Timestamp**       | When the match happened                             |
| **Tool**            | The AI tool in use                                  |
| **User**            | The developer involved                              |
| **Guardrails**      | The guardrail that matched                          |
| **Action**          | What was applied                                    |
| **Sensitive Count** | How many sensitive items were found in that request |

Ten rows to a page on both.

## Filters and controls

### Time range

The standard Analytics range selector applies.

### Filters

| Filter           | Values                              |
| ---------------- | ----------------------------------- |
| **Guardrails**   | **PII** and **Secrets**             |
| **Actions**      | **Audit** and **Block**             |
| **User**         | The people you can read             |
| **Tool Type**    | The AI tools your organization uses |
| **Department**   | Departments in your organization    |
| **Team Manager** | Managers in your organization       |

## What you can do

| Action                | Result                                                                                    |
| --------------------- | ----------------------------------------------------------------------------------------- |
| Select a Summary row  | Opens **Logs** filtered to that tool, guardrail and action, with the chips to show for it |
| Return to **Summary** | Undoes that drill-down. Filters you set by hand since the click stay put                  |
| Select a log row      | Opens the **match panel**, below                                                          |

### Match panel

Headed with the developer and the tool, then the request time.

| Section                       | Contents                                                                                                                                  |
| ----------------------------- | ----------------------------------------------------------------------------------------------------------------------------------------- |
| **Sensitive Information (N)** | Every sensitive item found. Where those values appear in the user prompt, **Highlight** marks them and *1 of N* steps through the matches |
| **User** · **Assistant**      | The prompt that matched and the reply, each expandable                                                                                    |
| **Tool Use** · **System**     | The tool calls and system prompt, where the request carried them                                                                          |

An Admin can also **Mark as False Positive** on a detection that was wrong, or **Report missing detection** when a guardrail should have fired and did not. Both tune future detection. An Analytics Viewer sees the panel read-only.

## Messages you may see

| Message                           | What it means                                                               |
| --------------------------------- | --------------------------------------------------------------------------- |
| `No Data Found` · `No Logs Found` | Nothing matches the current range and filters. Widen either                 |
| `Marked as not sensitive`         | The detection will no longer count as sensitive                             |
| `Couldn't update the detection`   | The mark was not saved, so the match still counts. Try again                |
| `Failed to mark as not sensitive` | The same failure, reported from the match panel. The detection still counts |
| `Couldn't submit the report`      | The missing-detection report was not sent. Try again                        |

## Related

| Page                                          | For                                                                  |
| --------------------------------------------- | -------------------------------------------------------------------- |
| [Policies — Data](/dashboard/policies/data)   | Authoring the guardrails reported here                               |
| [Policies — Usage](/dashboard/policies/usage) | Account Access refusals, which are reported on the policy's own page |
| [Logs](/dashboard/logs)                       | The full request behind a match                                      |