The creator, and authorised administrators, of an agent are able to see a de-identified list of conversations that users have had with that agent.
This ability is important because it allows:
- Teachers to better understand any misconceptions that students may have, to inform their teaching
- Teachers to monitor the quality of AI responses
- Monitoring of potential abuse
To access the conversation history, from Agent HQ find your agent and click on the View conversation history icon.
This will show a list of conversations, with the identity of the user(s) blanked out.
You can view the conversation itself by clicking on the speech bubble icon on the right.
If users have flagged an AI response, or rated it positively or negatively, this will be shown in the conversation history alongside any comments that the user left as feedback.
