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

# 2. Audit Log

> Walk clients through every retained AI prompt over the past 30 days and the policy violations Longwave caught and redacted on their behalf.

Open the tenant and go to **Reports → Audit Log**. This is the evidence layer:
every retained AI prompt across users and apps over the past 30 days, flagged
where Longwave detected a violation.

<Frame>
  <img src="https://mintcdn.com/longwave/-0YwkY__ulfCjmSG/images/audit-log.png?fit=max&auto=format&n=-0YwkY__ulfCjmSG&q=85&s=a863f2f446a8c7d76de879d175175868" alt="Longwave Audit Log for the Northwind tenant: list of retained prompts by user and AI app over the past 30 days, with an interaction detail panel showing a ChatGPT prompt from Michael Chen that contained an email address and US phone number, both redacted under the active AI policy" width="1024" height="723" data-path="images/audit-log.png" />
</Frame>

## What clients see in a review

Open the log together and walk through three things:

* **Who is using which AI app.** The list shows the user, their email, and the
  AI app they used (ChatGPT, Microsoft Copilot, Gemini, and so on).
* **What was actually sent.** Click any row to open the **Interaction**
  panel: the AI app, the user, when it happened, and the **prompt** as it was
  submitted.
* **What policy caught.** The **Violations** section lists each detector that
  fired (for example *Email Address*, *US Phone Number*) and shows the
  matching content was **Redacted** before it left the browser.

## The story to tell

That last part is what makes a customer renew. The user typed a customer
email and phone number into ChatGPT, and your policy quietly took them out.
The customer didn't lose their workflow, and the regulated data didn't leave
with the prompt.

A few of those, walked through by name, do more for retention than any chart.

## How to use it in a meeting

1. Open the tenant and go to **Reports → Audit Log**.
2. Filter or search to a few interesting users or apps.
3. Open one or two interactions where **Violations** fired. Walk the customer
   through what was redacted and why.
4. Tie it back to the **policy** they signed off on. This is the policy doing
   its job, with names attached.

<Info>
  Pick examples that map to the customer's real risks (PII, regulated
  identifiers, internal project names). One concrete redaction is more
  persuasive than a chart.
</Info>
