How Should a Power Virtual Agents Chatbot Handle Unrecognized Questions?
You deploy a Power Virtual Agents chatbot that integrates with Dynamics 365 Omnichannel for Customer Service. You observe that the chatbot is not able to recognize the questions asked by users. You need to ensure that the chatbot can respond to unrecognized questions. The solution must minimize administrative effort. What should you do?
Community Votes
100% of anonymous learners picked answer A. Votes are pick records left by other test-takers — they are not the verified answer.
Community Insight
It tests whether you know the Fallback system topic is automatically triggered when no authored topic matches an utterance — the trap is confusing it with the Escalate system topic or assuming you must author one topic per possible question.
When a Power Virtual Agents chatbot integrated with Dynamics 365 Omnichannel cannot recognize user questions, the Fallback system topic is the built-in mechanism that catches unmatched utterances. This page confirms that customizing the fallback topic (A) is the lowest-effort fix, while new topics, entities, and the Escalate topic do not address unmatched input.
Many learners pick the Escalate system topic (D) because both are system topics related to "something went wrong," but Escalate only fires on explicit escalation intent or a redirect action, so it never catches unrecognized questions.
Community Discussion (5 comments)
Comments & Corrections
No comments yet — spotted an error or have a note? Share it below.
Expert Analysis
Why the Answer Is Correct
Power Virtual Agents ships with system topics, and the Fallback topic is the one that fires automatically when the natural language understanding engine cannot match a user utterance to any authored topic. Editing that single topic — adding a message, a clarifying question, or an Omnichannel handoff — is the whole fix, so it satisfies the "minimize administrative effort" constraint. No new topic authoring, no trigger phrases to invent, and no per-question maintenance are required. This is also the Microsoft-recommended pattern for graceful degradation before escalating to a human agent in Dynamics 365 Omnichannel.Why the Other Options Are Wrong
Creating new topics (B) means trying to anticipate every unrecognized phrasing in advance, which is never complete and multiplies admin effort. Creating an entity (C) only extracts structured values (dates, order numbers) from utterances that already match a topic — an entity does not route or respond to unmatched input. Modifying the Escalate system topic (D) has no effect on recognition: Escalate triggers on explicit escalation intent or when another topic redirects to it, so unrecognized questions would never reach it.Community Comment Notes
Agnes202307 explains the core logic well, noting the fallback topic handles input "when the user's input doesn't match any specific topic" and does so "without requiring extensive administrative effort." MrEz adds a useful diagnostic caveat — it "depends what the reasons are, but fallback topics was my choice" — meaning you should first confirm the misses are genuinely unmatched utterances and not a broken knowledge source or trigger phrase. The remaining voters (33a7f9e, jkaur, Sweden2022) simply state "A is correct," so the community consensus aligns with the Fallback topic.Exam Strategy
When a PL-200 question mentions "unrecognized," "unmatched," or "unknown" user input, look for the Fallback system topic first — Microsoft frames it as the default catch-all. Keywords like "minimize administrative effort" are a strong hint toward configuring an existing system topic rather than authoring new ones.
Frequently Asked Questions
Why not just create new topics for the questions users ask?
New topics only match utterances you anticipated and authored trigger phrases for; unrecognized questions by definition have no matching topic, so the effort never ends.
Doesn't the Escalate system topic already handle unrecognized input?
No. Escalate fires on explicit escalation intent or when a topic redirects to it, so it never triggers on an unmatched utterance — only the Fallback topic does.