liveBook · Manning

Multi-intent utterance

As explained in Effective Conversational AI

A multi-intent utterance is conversational input that expresses two or more distinct user goals.

A multi-intent utterance expresses two or more distinct user goals in one conversational input. During first-pass review, flag it as a potential test candidate rather than over-analyzing it. section 4.4.2

Screen the utterance alongside other production-log categories:

  • Check whether it is unintelligible, unrelated to the domain, ambiguous, out of scope, or expresses an in-domain, in-scope goal. section 4.4.2
  • Make a quick determination; discard uncertain candidates for now, or mark recurring patterns for later review. section 4.4.2

How Effective Conversational AI covers it

How deep it goes
Defines the screening category and review behavior, but stops short of explaining how to decompose or resolve the goals.
Examples
Uses production-log utterances as candidate material and compares multiple intents with unintelligible, unrelated, ambiguous, and out-of-scope inputs.
What you can do after
Evaluate production utterances quickly and mark bundled goals as candidates for later testing or review.
How it connects
Connects multi-intent utterances with production logs, blind test sets, ambiguity, scope, and speech-to-text quality.
Cover of Effective Conversational AI
liveBookmanning.com
Read Effective Conversational AI on liveBook Get the book at manning.com

Related concepts

This explainer was generated by Manning's knowledge graph from the text of Effective Conversational AI. The section links point to the parts of the book it draws on.