What a step contains
Each step in the timeline has:- A screenshot — captured before the action was executed.
- The action — click, type, select, scroll, wait, navigate, complete, or abandon.
- Target details — for click/type/select, which element the agent targeted.
- The agent’s reasoning — what it was thinking when it chose the action.
Why the timeline matters
- Interprets ambiguous issues. When an issue says “the button was unclear,” the timeline shows you which button and what was on-screen when the persona saw it.
- Debugs abandonments. If the test abandoned early, the last few steps show you what tripped the agent up.
- Spots false positives. If an issue feels wrong, look at the screenshot — sometimes the agent misread the page and the issue doesn’t apply.
- Validates the run. Confirms Percio actually did the thing you wanted tested before you trust the score.
Action types
- Click — a click on a specific element.
- Type — text typed into an input.
- Select — a value chosen from a dropdown.
- Scroll — scrolling the viewport or a specific container.
- Wait — pausing for a page transition or async action to finish.
- Navigate — loading a new URL.
- Complete — the agent decided the scenario was done.
- Abandon — the agent decided the scenario was unreachable.
Reading a “stuck” run
If the timeline shows the agent clicking the same area repeatedly, or bouncing between the same few actions, the run was probably stuck. Common causes:- A modal or cookie banner the agent didn’t dismiss.
- A form that silently failed validation.
- A CTA that looked interactive but wasn’t (e.g. styled as a button but not clickable).
- A CAPTCHA or bot-detection interstitial.