Summarize sessions with PostHog AI

PostHog AI generates summaries of session recordings, so you can understand what happened in a session without watching the full replay. Each summary analyzes the events and actions within a recording and produces a concise description of what the user did.

How it works

A session summary typically includes:

• Key actions – pages visited, buttons clicked, forms submitted, and features used
• Behavioral patterns – repeated actions, navigation loops, hesitation, or rage clicks
• Errors encountered – JavaScript errors, failed network requests, or captured exceptions
• Session outcome – whether the user completed a goal, dropped off, or experienced issues

How summaries work

PostHog AI reads the event stream from a recording – every page view, click, input, scroll, error, and custom event – and produces a human-readable summary. It doesn't watch the video; it analyzes the structured event data, which means summaries are fast even for long sessions.

This makes it practical to summarize sessions in bulk. Instead of watching 20 recordings to spot a pattern, ask PostHog AI to summarize all of them and surface what's common.

Try it

Select a prompt to try it out in the PostHog app:

• Summarize the last 5 session replays where users dropped off during checkout
• What patterns do you see in sessions with rage clicks?
• Summarize sessions from users who signed up today
• What are users struggling with on the pricing page?

Summarize sessions across experiment variants

Session summaries are especially useful for understanding Experiment results. Metrics tell you which variant won, but summaries tell you why.

Ask PostHog AI to summarize replays for each variant to spot behavioral differences between control and test groups – like different navigation patterns, different points of confusion, or different feature usage.

• Summarize session replays for the control group in my checkout experiment
• What behavioral differences do you see between variant A and variant B?

Tips for better results

• Ask about specific behaviors – "What do users do after seeing the upgrade prompt?" gives you more targeted summaries than "Summarize recent sessions"
• Batch for patterns – Summarizing multiple sessions at once (e.g. "the last 10 sessions with errors") helps PostHog AI identify recurring issues
• Combine with user properties – "Summarize sessions from enterprise users who churned" narrows the focus to the most relevant recordings
• Use with experiments – Comparing summaries across variants reveals behavioral differences that raw metrics might miss
• Ask follow-up questions – After getting a summary, ask "What's the most common point of drop-off?" or "Were there any errors?" to drill deeper

Get started

To use session summaries, set up PostHog AI. Session summaries require you to enable Session Replay in your project.