Agentic AI Reference Architecture

Human Oversight · Software component

Failure-Triggered Feedback Prompt

Software componentHuman OversightExperience & Human Oversightarc:FailureTriggeredFeedbackPrompt

A feedback collector that proactively solicits a reason when behavioural signals such as abandonment or rapid escalation indicate an unhelpful response.

Responsibility. Converts detected failed interactions into targeted feedback opportunities.

Also known as: Contextual failure prompt

Variant of Feedback Collector abstract

When to choose. Choose when training data should focus precisely on current weaknesses rather than random optional feedback.

specializesis triggered byis target of alternativeTois target of alternativeToalternative toFeedback Collector: specializesFeedback CollectorBehavioral Signal Tracker: is triggered byBehavioral Signal TrackerDeferred Feedback Survey: is target of alternativeToDeferred Feedback SurveyBinary Rating Feedback Prompt: is target of alternativeToBinary Rating Feedback P…Structured Correction Feedback Form: alternative toStructured Correction Fe…
Direct neighbourhood (hover for relationship types)

Relationships

is triggered by dynamic

alternative to variability

Design guidance

Classification

Patterns
Targeted solicitation at failure points
Quality attributes
Functional suitability: correctness and validity (ISO/IEC 25010 | NIST AI RMF: valid)
Risks mitigated
Feedback samples dominated by successful cases with little improvement signal

Sources

  1. Ch10.1: T. Nguyen, "Conversational UI," in Mastering Agentic AI Systems: Guide for the NVIDIA NCP-AAI Exam, 1st ed. 2026, ch. 10.1. ISBN: 9798244538229.