One task, one history
A run is one execution of a task. You can inspect its status, duration and recorded steps. A log is a record of a step, such as a tool call or error.
Insights and logs
See what your agent actually did.
Review the steps taken for each task, the tools used and where something went wrong. This is observability: visibility into the work behind an answer.
How it works · example
Why was a customer question left unresolved?
The request
A customer asks about an order. The agent tries to look up its status.
What you see
The task history shows a failed tool call. The connection to the order system returned an error.
Your next step
Check the connection and test the question again. This helps you distinguish a connection problem from an instruction problem.
Example using a connected order system. The details available depend on the tool and the events recorded.
For administrators who need to make changes
Use this information to repair a connection or adjust instructions. Logs show what happened; they do not fix a problem on their own.
Want to see how to inspect a completed task? We can demonstrate it with a process from your business.
Request a demo