I have found this update especially useful when reviewing experiment outcomes that show a clear pattern in the numbers but do not fully explain the reason behind it.
For the exact release notes and rollout context behind this update, you can refer to PostHog’s official changelog.
Turning Experiment Results into Usable Insight
When I’m reviewing experiment results, I usually do two things: confirm the deltas/statistical outcome, then jump into session recordings to understand what people actually did. The problem is volume: I don’t want to watch a pile of replays just to find the pattern.
Now, inside experiment results, there’s an option to summarize session replays tied to the experiment. The way I’m using it is straightforward: I open the experiment, click a variant in the results visualization to open the detail panel, and from there I either view recordings or click “Summarize session replays.” PostHog AI then analyzes recordings across variants and surfaces behavior patterns, specifically calling out differences in how users interact with each variant.
What I like about this is that it’s not replacing the stats—it’s giving me the “why” layer right where I’m already making the decision. I still validate by opening a few representative recordings, but now I’m doing that with a hypothesis in mind instead of hunting blindly.
