Skip to content
anythingai.space
Conversational AIintermediate12 min · 9 steps

Dialogue Policy and NLG

How a classical bot decides what to do next from its dialogue state, and how NLG turns that decision into a sentence with templates, response selection or a neural model.

Dialogue statecity ✓ · check_in ✗Policystate → action?ASK(check_in)CONFIRMEXECUTE bookingASK_CLARIFICATIONESCALATE
datarequestsignalresponsetoken
1 / 9
ProblemStep 1 of 9

The state is known. Now what?

After NLU and state tracking, the bot knows: intent BOOK_HOTEL, city Paris, check-in date missing. It still has to choose one thing to do from its action space: ask for the date, confirm what it has, book, ask a clarifying question, or hand off to a human.

That choice is the job of the dialogue policy. A good policy asks only what's needed, recovers from misunderstandings, and moves the user towards a finished task.

Lesson 3 of 8 in Conversational AI