The answer
The Users view begins with observed users, repeat users, users with issues, cost per chat, and average execution time. A population map places analyzed session count against the share of sessions without issues.
Activation depth separates first-session users, users with two to four sessions, and users with five or more. The action queue then surfaces useful, explainable groups such as new users and power users to learn from.
Need a business-specific comparison? Groups can compare sessions by customer, model, environment, or another property present in your traces.
Which behavioral cohorts are worth starting with?
- New users against returning users.
- Users who completed their task against users who abandoned it.
- Users who hit an error or a loop in their last session.
- Your most active users, who often show the next feature people will ask for.
- Users by channel or AI client, since behavior differs between them.
- Users by plan or account size, to connect behavior with revenue.
Published