liveBook · Manning

Out-of-scope utterance

As explained in Effective Conversational AI

An out-of-scope utterance expresses a clear goal relevant to a domain but not supported by the current conversational solution.

A user may express a clear goal that belongs to the domain, yet the current conversational solution does not support it. Treat that input as an out-of-scope utterance rather than as an unrelated or unintelligible request. section 4.4.2

During a first-pass review, distinguish out-of-scope utterances from nearby categories:

  • Unintelligible utterances cannot be understood reliably and may result from typing errors, speech transcription, or background noise. section 4.4.2
  • Completely unrelated utterances are understandable but do not concern the domain. section 4.4.2
  • Out-of-scope utterances are domain-related, but their goal is not supported by the current solution. section 4.4.2
  • Reviewers should make a quick initial determination, set conflicted cases aside, and avoid over-analyzing individual utterances. section 4.4.2

How Effective Conversational AI covers it

How deep it goes
Provides a first-pass classification checklist and review guidance, without detailing downstream handling or implementation.
Examples
Uses production-log utterances, screening criteria, and blind test sets as review artifacts.
What it assumes
Assumes a defined domain and conversational scope, then teaches reviewers to apply those boundaries to candidate utterances.
What you can do after
Evaluate utterances consistently by separating unsupported domain goals from unintelligible, unrelated, ambiguous, or multi-intent inputs.
How it connects
Connects domain relevance with intents, production logs, and blind test sets.

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.