Ask

Ask lets you talk to your own data. Type a question in plain language and get an answer from your organisation's Agile Analytics data, with charts and links to the pages behind the numbers. It is the same assistant that answers when you mention the bot in Slack.

Asking a question

The chat opens with a few suggestions, such as Which areas need attention right now?, How did our lead time develop this quarter?, Who received the most kudos this month? or What is <a colleague> working on?. Press Tab to take a suggestion, or type your own. Follow-up questions keep the context of the chat; New chat starts over. An answer can take up to half a minute.

The assistant can look at:

• the engineering health score and its areas, now and over time;
• the DORA metrics, sprint predictability and flow, error budgets, booked hours and kudos;
• AI adoption and the AI leaderboard, where the plan and the settings allow it;
• your teams, and what a person is working on: open and merged merge requests, commits, Jira work, hours and kudos;
• the signals of the last weeks.

It only reads data your organisation already has in Agile Analytics. It cannot change anything, with one exception: when you ask it to put a team together, it drafts one.

Draft teams

Ask something like make a team of the people who work on the payments repositories. The answer shows a draft team with its members and a Create team button. People with write or admin rights can create it; add repositories, services and Jira boards afterwards on the Teams page.

Status

The strip above the chat shows whether the assistant is ready, whether the Slack bot is connected, how many questions were answered in Slack and here, how many kudos were given recently, and how many Slack members are linked. The Slack card lists the channels the bot is in and opens a direct message with it in Slack. If the bot is in no channel yet, invite it: see Slack app.

About the answers

Answers are written by an AI model (Claude by Anthropic) from your organisation's data. They can be inaccurate or incomplete, so check important numbers on their page. Questions are sent to Anthropic to write the answer and are not used to train models. Each organisation has a daily limit of questions.