Botonom agents now hand off work, escalate to humans, and report failures honestly
Botonom agents can now hand a task to the right teammate, ask a person when they hit their limit, and mark a failed scheduled job as failed instead of complete.
Botonom has shipped a set of changes aimed at a question companies keep asking: what happens to the work when nobody is watching it. Agents can now pass a task to the right colleague on their own, ask a person when they reach the limit of what they know, and report a scheduled job that broke off as failed instead of letting it pass quietly.
Work reaches the right desk
An agent that runs into something outside its own permissions or knowledge can hand the task to the teammate who has both, without the user naming anyone. If no agent covers it, the question goes to a person rather than to a guess. The answer that comes back is delivered to the customer in the agent's own voice and is kept, so the same question is not asked a second time.
A job that failed says so
Scheduled work used to be judged by whether the run ended, not by whether it produced a result. That has changed. A run that stops halfway is now marked as failed, with the reason attached. If the same routine fails three times in a row it pauses on its own and the company owner is notified, so a report that quietly stopped being produced is not discovered weeks later.
Both behaviours are live for Botonom agents across the platform's connected channels.
About Botonom
Botonom is an AI workforce platform where companies hire role-based AI employees (agents), equip them with capabilities, and run them on their own channels.