AI-102 Plan and manage an Azure AI solution Practice Question
A developer is building a chatbot using Azure Bot Service and Language Understanding (LUIS). The bot needs to handle multiple intents, including 'BookFlight', 'CancelFlight', and 'CheckWeather'. During testing, the bot frequently confuses 'BookFlight' and 'CancelFlight' intents. What is the most effective way to improve intent classification accuracy?
⚠ Common exam trap
It's easy for candidates to confuse confidence thresholds with model improvement, thinking that raising the threshold will fix misclassifications, when in reality it only masks the problem by rejecting more utterances instead of improving the model's discriminative power.
Answer choices
Why each option matters
Answer the question above first, then reveal the full breakdown to understand why each option is right or wrong.
Correct answer & explanation
✓
Add more varied training utterances for 'BookFlight' and 'CancelFlight' intents.
Adding more varied training utterances for the 'BookFlight' and 'CancelFlight' intents directly addresses the root cause of confusion: insufficient or overlapping training data. LUIS relies on diverse utterance patterns to distinguish between semantically similar intents; increasing the quantity and variety of labeled examples improves the model's ability to learn discriminative features, thereby boosting classification accuracy.
Answer analysis
Option-by-option breakdown
For each option: why learners choose it and why it is or isn't the right answer here.
- ✗
Reduce the number of intents by merging similar ones.
Why it's wrong here
Merging intents removes the distinction the bot must resolve, so BookFlight and CancelFlight become one label and the required behaviour is lost. Consolidation is tempting when intents genuinely overlap in meaning, but here they are semantically opposite and must remain separate.
- ✗
Increase the confidence threshold for intent predictions.
Why it's wrong here
Raising the threshold only suppresses low-confidence predictions; it cannot separate utterances whose trained features overlap, so confusion persists or utterances fall to None. Threshold tuning is tempting for filtering uncertain matches, and would suit rejecting out-of-scope input, not correcting a mislabelled intent boundary.
- ✗
Add more entities to the utterances.
Why it's wrong here
Entities label extracted data within an utterance; they do not adjust which intent the classifier assigns, so the BookFlight/CancelFlight boundary stays confused. Entity enrichment is tempting because it improves slot filling, and would be right when the intent is correct but key values are missed.
- ✓
Add more varied training utterances for 'BookFlight' and 'CancelFlight' intents.
Why this is correct
LUIS learns intent boundaries from labelled utterances, so adding varied, representative examples for BookFlight and CancelFlight sharpens the model's discrimination between them. This addresses the root cause: insufficient or overlapping training data for the confused intents.
Go deeper
Related to this question
About these practice questions
This AI-102 question is part of Courseiva's 761-question bank — original exam-style content with full explanations and wrong-answer analysis, never real exam questions or exam dumps. Learn why practice questions differ from exam dumps →
JA
Written by Johnson Ajibi, MSc IT Security
Senior Network & Security Engineer · founder of Courseiva
This AI-102 practice question is part of Courseiva's free Microsoft certification practice question bank. Courseiva provides original exam-style practice questions with explanations, topic-based practice, mock exams, readiness tracking, and study analytics to help learners prepare for the AI-102 exam.