> For the complete documentation index, see [llms.txt](https://docs.cortex.io/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://docs.cortex.io/improve/opex-review/using-opex-ai-chat.md).

# Using Cortex AI in OpEx Review

Cortex AI is Cortex's AI-powered chat assistant. It answers questions about your engineering data in plain language, using the same data that powers your Scorecards, dashboards, and reports. 

During an OpEx meeting, leaders often ask questions that no one in the room can answer on the spot, like which team is driving a spike or how a metric compares to last quarter. Normally, those questions become post-meeting follow-ups, and the discussion loses momentum. Cortex AI lets you ask the question and get an answer without leaving the report.

{% hint style="info" %}

#### Known limitations

Cortex AI chat is in Research Preview and available to a limited group of customers. To request access, reach out to your Customer Success Manager.

* Responses can be slow, especially for questions spanning many teams or entities.
* Not all Cortex data is reachable from chat yet, so it can't answer every question about your organization.
* It can't take action on your behalf. It won't create or edit Scorecards, Initiatives, Dashboards, or Workflows, and it won't change the report's content or filters.

We're actively working on improvements. Share feedback with your account team or through the in-product feedback controls.
{% endhint %}

## Accessing Cortex AI

You can chat with Cortex AI from the summary section on the **Executive briefing** tab or from a metric card. 

`Admins` and users with the `Enable AI Chat` permission can access Cortex AI.

**To access Cortex AI via the summary section**:

1. From the main sidebar, select **OpEx Review**.
2. From the **Executive briefing** tab, locate the summary section on the left side of the page.
3. At the bottom of the summary, do one of the following:
   1. Select a pre-populated question.
   2. Enter your question in the **Ask Cortex AI** field, then press **Enter**.<br>

      <div align="left" data-with-frame="true"><figure><img src="https://826863033-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FJW7pYRxS4dHS3Hv6wxve%2Fuploads%2Fo06KmNUv9Qyp5XGeelxA%2Fopex-cortex-ai-summary.png?alt=media&amp;token=eb7fd4f6-0496-4139-b75d-6460acc84072" alt="Cortex AI options at the bottom of the summary, in OpEx Review." width="375"><figcaption></figcaption></figure></div>

**To access Cortex AI via a metric card**:

1. From the main sidebar, select **OpEx Review**.
2. From the **Executive briefing** tab, locate the metric card you want to dig into.
3. At the bottom of the card, click **Ask Cortex AI**. The side panel opens with the metric summary pre-populated.<br>

   <div align="left" data-with-frame="true"><figure><img src="https://826863033-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FJW7pYRxS4dHS3Hv6wxve%2Fuploads%2Fz2vs14YZHijQqqZKdbD7%2Fopex-ask-cortex-ai.png?alt=media&amp;token=bc6dec7c-9d00-4bfc-9341-54ee718b921f" alt="&#x27;Ask Cortex AI&#x27;, highlighted, at the bottom of a metric card." width="375"><figcaption></figcaption></figure></div>
4. Do one of the following:
   1. Select a pre-populated question.
   2. Enter your question in the **Ask Cortex AI** field, then press **Enter**.

Cortex AI has access to the same data behind the insight, plus the ability to query further breakdowns, such as by team, entity, or time period, so it isn't limited to just restating the insight you started from.

Each chat is tied back to the metric it came from. If you open the panel again later, or view it as a full chat from the Cortex AI section in the main sidebar, you'll see a note showing which report and metric it was forked from, with a link back to that report.

## What you can ask

Cortex AI is built for follow-up questions on whatever metric or insight you're looking at, for example:

* ***Which teams or services are driving this increase?***
* ***How does this compare to last quarter?***
* ***What changed in the last two weeks?***
* Why a specific team's metric moved (for example, ***Why did Platform Engineering's deploy frequency drop?***)
* How two teams or time periods compare
* What a metric or threshold means, useful for less technical leaders in the room

See the [Cortex AI prompt library](/get-started/cortex-ai-assistant/library.md) for a full list of prompt patterns organized by use case and role.

## What to expect from a response

Answers are intended to be discussed in your OpEx meeting. Responses are typically two to four sentences, or a short bulleted list. Responses are designed to:

* Call out specific teams or services by name where relevant
* Suggest a next step when there is one
* Link to a relevant data source, such as an entity page, filtered view, or Data Explorer, if it helps to dig in further

Cortex checks any metric it surfaces against the query engine, but as with any AI-generated summary, it's worth double-checking a number before it drives a decision.