Skip to main content

​
What This Covers

This page explains how to test your agent before publishing. Testing catches problems before your agent runs automatically.

​
Why Test

  • Catch errors early — Find problems before they affect real workflows
  • Verify output format — Make sure results look right
  • Confirm destination — Check that data goes where you expect
  • Build confidence — Know your agent does what you intended

​
Two Testing Methods

When to use: Any agent. Best for scheduled agents or when you want to test immediately.How it works: Click Play, and your agent runs right now. You watch each step execute in real time.
1

Click Play

Find the Play button (▶️) in the topbar on the right side.
2

Watch execution

See each step as it happens: data fetched, actions taken, results generated.
3

Check output

Verify the final result. Did the email arrive? Is the spreadsheet row correct? Is the Slack message formatted right?
4

Review run details

Click on any step to see exactly what data was used and what was returned.

​
What to Verify

Data retrieved correctly — Does the agent get the right information?
Processing accurate — Are summaries, calculations, or transformations correct?
Output format right — Does the result look how you expected?
Delivered to correct place — Did it go to the right email, channel, or spreadsheet?
Handles edge cases — What happens if data is missing or unexpected?

​
Viewing Run History

Every run is saved so you can review what happened. Where to find it: Click the History button (🕐 clock icon) in the topbar.

​
What You’ll See

ElementMeaning
Run number (#1, #2, etc.)Order of execution, newest first
TimeWhen the run started
Triggered badgeRun was automatic (from schedule or event)
Blue highlightYou haven’t viewed this run yet

​
Reviewing a Run

Click any run to see:
  • Each step the agent took
  • Data retrieved from apps
  • Actions performed
  • Final output delivered
  • Any errors that occurred
New background runs (from schedules or triggers) show a notification badge on the History button. Click to review them.

​
Common Issues

SymptomLikely causeWhat to check
Agent doesn’t find dataWrong time range or filterCheck your date ranges and search criteria
Output goes to wrong placeIncorrect destination in instructionsVerify email addresses, channel names, spreadsheet links
Format looks wrongInstructions unclear about formatBe more specific about how output should look
Agent errors on a stepApp not connected or permission issueClick Accounts in topbar to reconnect
Trigger not detectedTrigger not published or wrong configEnsure event trigger shows green “Live” status
Results are emptyNo data matches criteriaBroaden your search or check that source data exists
Test button missingNo event trigger configuredTest button only appears for event-triggered agents

​
After Testing

Once testing confirms your agent works:
1

Save the version

Your agent saves automatically whenever you make changes.
2

Set your trigger

Configure when the agent runs: schedule (daily, weekly) or event trigger (new email, etc.).
3

Activate

  • For schedules: Click Enable in the schedule settings. The schedule activates immediately.
  • For event triggers: Click Publish in the topbar. Status changes from Draft (amber) to Live (green).
4

Monitor first runs

Check the first few automatic runs in History (🕐) to confirm everything works.
Test after changes. Whenever you modify your agent’s instructions or app connections, test again before relying on automatic runs.

​
Testing Checklist

Run a quick test with Play button
Check that output appears in the correct destination
Verify output format matches expectations
For event-triggered agents, do a trigger test with real event
Review any errors in History (🕐)
Enable schedule or publish event trigger only after successful test
Next: Triggers — Schedule and Event-Based Automation →