Feedback

Rating the AI’s summaries tells your administrators and the Flexivity AI team whether they are giving you what you need, and where to adjust them.

Rating a summary

Every summary in the Context section of the Flexivity panel has two buttons:

• 👍 Helpful: the summary gave you the context you needed.
• 👎 Not helpful: it missed something important, got something wrong, or was too long to be useful.

One click records it, and the panel says Thanks!. There is no form and no comment. You rate each summary once.

What happens next

• Your rating is combined with your team’s, so patterns show: for example, summaries of one kind of ticket that are often rated down.
• Your administrators see the trends on the User Feedback tab of Analytics in the Flexivity admin console, and can change how summaries are written: their instructions and tone, on AI Configuration.
• Nothing changes straight away. A rating does not rewrite the summary you rated, and does not train the AI. It is evidence for the people who tune it.

Who sees your rating

Ratings are anonymous. No agent ID, email or name is recorded with them. Your administrators and the Flexivity AI team see trends, not who rated what, and other agents do not see your ratings.

Other feedback the AI gets

You do not rate the AI’s other work. It is measured from what you do instead:

• Attribute values you correct in Context are recorded, so your administrators can see how often the AI’s values are changed. See AI Context.
• Draft replies you use, with Use this reply, or reject, with Not useful, are recorded too. See Progressive Autonomy.

Rating well

• Rate for usefulness, not perfection. If the summary gave you what you needed, that is a thumbs up.
• Don’t default to thumbs up. A thumbs down on a summary that missed something is the most useful rating there is.
• Go with your first reaction. A quick, honest rating beats a considered one you never give.