Copyright (c) 2026 MindMesh Academy. All rights reserved. This content is proprietary and may not be reproduced or distributed without permission.
8.2. Common Exam Traps and Distinctions
The AI-102 exam includes several common misconceptions that trip up candidates. Review these traps carefully—understanding why the wrong answers are wrong is as important as knowing the correct answer.
Service Selection Traps
| Trap | Correct Understanding |
|---|---|
| OCR is NLP | OCR is Computer Vision (input is image) |
| Document Intelligence = OCR | Document Intelligence extracts structured data, not just text |
| DALL-E can analyze images | DALL-E generates images; use GPT-4o Vision for analysis |
| Question answering handles dynamic data | Only works with static information |
Configuration Traps
| Trap | Correct Understanding |
|---|---|
| Lower temperature = more creative | Lower temperature = more deterministic |
| Content being blocked? Change categories | Adjust severity levels, not categories |
| Substitution errors in STT | Add custom vocabulary (names, products) |
| BLEU 30 is high quality | 40-59 is high quality |
| Special characters OK in synonyms | Special characters (#, @, %) are not allowed |
Security Traps
| Trap | Correct Understanding |
|---|---|
| Containers work offline | Still need billing endpoint |
| API keys are secure enough | Managed identity is most secure |
| All networks is fine for production | Use Private endpoint for Azure-only access |
API/Implementation Traps
| Trap | Correct Understanding |
|---|---|
| DALL-E needs size in body | Only prompt is required in body |
| Translator needs one header | Needs both Ocp-Apim-Subscription-Key AND Ocp-Apim-Subscription-Region |
| Password-protected PDF size issue | Password protection always fails regardless of size |
Written byAlvin Varughese
Founder•15 professional certifications