WEBVTT

NOTE AI Engineering Visualised: https://understandingdata.com/ai-engineering-visualised/few-shot-prompting/

1
00:00:00.250 --> 00:00:12.208
Routing a support ticket looks easy, but the right queue often hangs on house rules, like single sign-on being a plan feature that belongs to the account team, and no model can guess those.

2
00:00:12.208 --> 00:00:24.141
With no examples, the model gets five of the eight tickets right, sends single sign-on to feature requests in all ten runs, and flips the seats ticket between billing and account.

3
00:00:24.141 --> 00:00:33.845
Three examples picked from the edges of the rules lift it to seven of eight, and the SAML example beside it fixes single sign-on in all ten runs.

4
00:00:33.845 --> 00:00:46.766
Three near-identical refund examples add little: they steady the seats ticket, but single sign-on and the audit log stay wrong in every run, at twice the tokens of no examples.

5
00:00:46.766 --> 00:00:59.376
Here is the failure mode: add one example that calls a double charge after a crash a bug, and that ticket goes to bug in nine runs of ten, even with the three good examples still there.

6
00:00:59.376 --> 00:01:13.876
The part people miss is that examples teach neighbours, not rules, so an extra example about deleted projects fixes the audit log in only three runs of ten, and knocks the seats ticket down to two.

7
00:01:13.876 --> 00:01:25.096
So write the rule down and let examples show its edges: the house rules plus the same three examples route all eight tickets correctly in every one of eighty runs.