---
title: "DX AI chat"
canonical_url: "https://docs.getdx.com/dx-ai-chat/"
md_url: "https://docs.getdx.com/dx-ai-chat.md"
last_updated: "2026-05-22"
---

# DX AI chat
DX AI chat is a conversational interface for exploring your team's engineering data. Accessible from the DX AI sidebar, it lets you ask questions about your system data and snapshot data, dig into trends, and generate visualizations.

## Starting a conversation

There are two ways to start a chat:

- **New conversation** — Start a fresh chat from the DX AI sidebar. The AI knows which teams you manage and can answer questions about any of them.
- **From a Pulse alert** — Open any [Pulse](https://docs.getdx.com/pulse/) alert and start asking questions. The AI has full context of that alert, including the metrics, trends, and time period, so you can jump straight into follow-ups like "Why did cycle time spike?" or "Which repos drove the throughput drop?"

When starting a new conversation, you'll see suggested prompts to get started, such as:

- "Where should we focus?"
- "How many deployments were there last week?"
- "What are themes from the latest snapshot?"

## What the chat can do

- **Explore system metrics** — Answer questions about your team's velocity, quality, focus, and contributor activity using data from your connected tools
- **Surface snapshot insights** — Pull in themes, sentiment, and feedback from your team's DX snapshots
- **Visualize data** — Generate charts and visualizations backed by your system metrics and snapshot data
- **Compare trends** — Compare performance across teams or time periods
- **Diagnose changes** — Help you understand what's driving a metric change

## DX AI chat vs. Data Studio AI assistant

DX AI chat and the [Data Studio](https://docs.getdx.com/data-studio/) AI assistant both let you explore your engineering data with AI, but they're designed for different use cases.

**DX AI chat** is built for engineering managers who want quick, conversational answers about their teams. You ask questions in plain language — like "Why did cycle time spike last week?" — and the AI returns insights, charts, and summaries drawn from your system data and snapshot data. You never see or write SQL.

**Data Studio AI assistant** is built for anyone who wants to query the full DX database directly. It generates SQL from natural language requests, and you can see, edit, and run the resulting queries. It has access to the full data schema and supports features like template variables, saved queries, and datafeeds.

## Scope and limitations

- **Defaults to your teams** — When you ask a question without specifying a team, the AI answers using data from the teams you're a member of or manage. If you need a broader view — for example, across an entire department or organization — just ask, and the AI will pull in that data.
- **Read-only** — Chat cannot modify your settings, create alerts, or take any action on your behalf.
- **Engineering metrics only** — DX AI chat is purpose-built for team insights. It won't answer general questions outside that scope.

## IC-level data access

Administrators can control who can view individual contributor (IC) data in AI chat sessions using the **DX AI chat** setting under [Individual metrics visibility](https://docs.getdx.com/ic-metrics/). This setting applies to both DX AI chat and the [Data Studio](https://docs.getdx.com/data-studio/) AI assistant.

When a user without permission to view IC-level data asks a question that would surface it, the AI will respond with aggregate data only or indicate that IC-level data is restricted.
---

## Sitemap

[Overview of all docs pages](/llms.txt)