Error Handling & Failure Modes
Error handling describes what happens when AI Employees encounter problems while running skills or using integrations. Understanding these failure modes helps you troubleshoot issues and know what to expect when something goes wrong.
Incomplete or Missing Data
What Happens
When a skill runs but the trigger event is missing required data, the skill fails gracefully. For example, if a skill is supposed to create a task when an email arrives, but the email doesn't include a subject line, the skill cannot proceed.
How It's Logged
A failed skill run always tells you why it failed. The log indicates:
Which step the skill failed on
What data was missing or incomplete
A general error code (like "400" for a bad request)
What You Can Do
Review the error in the skill's run history. The error message will tell you exactly what went wrong. You can then adjust the skill's configuration to handle missing data differently, or add validation rules to prevent the skill from running when required data is missing.
Example: A skill is set up to "create a task when an email arrives." An email comes in but has no subject line. The skill fails and logs: "Failed at step 'Create Motion Task' — missing required field: task title." You can then update the skill to use a default title if the email subject is missing.
Permission Failures
What Happens
When an AI Employee attempts an action that the context user doesn't have permission to perform, the action is blocked immediately.
How It's Logged
The skill logs the permission failure with details about:
What action was attempted
Why it was blocked (permission denied)
Which user's context was used
Which workspace or project the action was attempted on
What You Can Do
Review the error and check the context user's permissions. If the context user needs to perform this action, you may need to:
Upgrade the context user's role (from Member to Admin, if needed)
Grant the context user access to the workspace or project
Adjust the skill to perform a different action that the context user has permission to do
Example: A skill tries to delete a project, but the context user is a Member (not an Admin). Members cannot delete projects, so the action fails. The log shows: "Permission denied — only Admins can delete projects." You can either upgrade the user to Admin or change the skill to archive the project instead (which Members can do).
Integration Failures
What Happens
When an external service (like Gmail or Salesforce) is unavailable or the connection has expired, the skill cannot complete actions that depend on that service.
How It's Logged
The skill logs the integration failure with details about:
Which external service failed (Gmail, Salesforce, etc.)
Why it failed (service unavailable, connection expired, authentication failed, etc.)
When it happened
Whether Motion will retry the action
Retry Policy
Motion automatically retries failed integrations according to a retry policy. If the service is temporarily unavailable, Motion may retry the action a few times before giving up.
What You Can Do
If an integration failure persists:
Check if the external service is having issues (check their status page)
Verify the connection is still valid (it may have expired)
Reconnect the integration by going to your workspace settings
Update any skills that were using that connection
Re-run the skill once the connection is restored
Example: A skill tries to send an email via Gmail, but the Gmail connection has expired (maybe you changed your Gmail password). The skill fails and logs: "Gmail connection expired — authentication failed." You reconnect Gmail in your workspace settings, and then re-run the skill.
Data Type Mismatches
What Happens
When trigger output data doesn't match what the skill expects, the skill fails with a clear error message. For example, if a trigger outputs a date in one format but the skill expects a different format, the skill cannot proceed.
How It's Logged
The skill logs the type mismatch with details about:
Which step failed
What data type was expected
What data type was actually received
The specific field or value that caused the problem
What You Can Do
Review the error message to understand the mismatch. You can then:
Adjust the trigger output mapping to convert the data to the correct format
Add a data transformation step in the skill to convert the data
Modify the skill to accept the data type that the trigger is actually providing
Example: A Motion trigger outputs a date as "2025-11-03" but the skill expects "November 3, 2025." The skill fails and logs: "Data type mismatch — expected date format 'Month Day, Year' but received 'YYYY-MM-DD'." You can add a transformation step to convert the date format before the skill uses it.
No Silent Failures
Important Principle
AI Employees never silently fail. Every failure is logged and visible to you.
This means:
If a skill fails, you'll see the error in the run history
The error message will tell you why it failed
You can review the log to understand what went wrong
You can take action to fix the problem
There are no hidden failures or skills that stop working without you knowing about it.
Where to Find Failure Logs
You can view failure logs in:
Chat history — For manual mode actions, errors appear in the chat
Skill run history — For autonomous mode skills, errors appear in the run history
Workspace activity feed — Some errors may also appear in the workspace activity log
Example: A skill is set up to automatically update Salesforce records every Friday at 5 PM. One Friday, the Salesforce connection is revoked. The skill fails and logs the error. You see the failure in the skill's run history and can reconnect Salesforce to fix it.
Last updated
Was this helpful?