# What made you change when your support agent hands a task to a person? — retained record

> Retained reddit record from reddit:AI_Agents.

Canonical: https://highsignal.app/data/records/86219eabfa2477f4

[← sources](https://highsignal.app/data)


source: [reddit](https://highsignal.app/data/reddit) (reddit:AI_Agents)  published 2026-11-18 01:32 ingested 2026-11-19 02:35

[original source ↗](https://www.reddit.com/r/AI_Agents/comments/1xz2gou/what_made_you_change_when_your_support_agent/)

## retained text



```
For anyone running a support agent, I’m interested in how you decide which tasks it should finish itself and which it should hand over.

Have you changed that boundary after reviewing real conversations? What prompted it: wrong answers, customers returning, people correcting completed tasks, or something else?

How did you check whether the new rule helped rather than simply sending more work to people?

I’m a founder researching workflow evaluation. Our own support example is a lab test, so I’m looking to understand how this decision works in production.
```



document fetched 2026-11-19 00:19 · canonical https://reddit.com/r/AI_Agents/comments/1xz2gou/what_made_you_change_when_your_support_agent
