Describe features of generative AI workloads on Azure →mediumMultiple ChoiceObjective-mapped

AI-900 Practice Question: Describe features of generative AI workloads on Azure

A developer is using Azure OpenAI Service to classify customer support tickets into categories such as 'Billing', 'Technical Issue', and 'Account Management'. The developer provides three labeled examples for each category in the prompt to improve the model's accuracy. What technique is the developer applying?

⚠ Common exam trap

It's easy for candidates to confuse few-shot learning with fine-tuning, assuming that any use of examples to improve accuracy must involve retraining the model, but few-shot learning does not modify model weights—it only uses examples in the prompt.

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

✓

Few-shot learning

Few-shot learning is the correct technique because the developer is providing a small number of labeled examples (three per category) directly in the prompt to guide the model's output without updating the model's weights. This approach leverages the model's in-context learning ability, where the examples act as a pattern for the model to follow when classifying new tickets.

Answer analysis

Option-by-option breakdown

For each option: why learners choose it and why it is or isn't the right answer here.

✗
Fine-tuning
Why it's wrong here
Fine-tuning would involve taking the base model and retraining it on a curated dataset of support tickets with their correct categories, updating the model's parameters through a training loop and potentially using GPU resources. The developer's approach does not include any retraining or parameter updates; it only modifies the input prompt. Azure OpenAI offers fine-tuning as a separate capability, but this scenario clearly does not use it.
When this WOULD be correct
A question where the developer trains a custom model on thousands of labeled support tickets to improve classification accuracy, and the model's weights are updated via additional training. This would be fine-tuning.
✓
Few-shot learning
Why this is correct
Few-shot learning is a form of in-context learning where the prompt includes a small number of labeled examples (e.g., ticket text and its category) to condition the model's behavior without updating its weights. The model infers the classification pattern from these examples and applies it to new tickets. This approach is efficient because it avoids retraining and works directly in the Azure OpenAI service at inference time.
✗
Prompt engineering
Why it's wrong here
Prompt engineering is the broad discipline of crafting the input text to get the desired behavior from a language model, and it includes many strategies like instruction formatting, chain-of-thought reasoning, and role playing. Providing a few labeled examples is just one specific prompt-engineering technique, and that technique has its own name: few-shot learning. Marking 'prompt engineering' alone is too general because it does not identify the concrete method used in the prompt.
When this WOULD be correct
A developer is crafting a prompt to instruct the model to output responses in a specific JSON format. The developer adjusts the wording and structure of the prompt to improve the model's adherence to the format. This scenario would test prompt engineering.
✗
Retrieval-augmented generation
Why it's wrong here
Retrieval-augmented generation (RAG) extends a model by connecting it to an external knowledge base, where relevant documents or snippets are retrieved (for example, via vector search) and inserted into the prompt to ground the answer. Here there is no retrieval step, external index, or dynamic data injection; the developer simply hardcodes a few examples into the prompt. Because no information is fetched from an outside store, RAG is not what is happening in this classification task.
When this WOULD be correct
A developer needs to answer customer questions using a large internal knowledge base. They use Azure OpenAI Service to retrieve relevant documents from the knowledge base and include them in the prompt to generate accurate answers. This scenario describes retrieval-augmented generation.

Option-by-option analysis

Why each answer is right or wrong

Understanding why wrong answers are wrong — and when they would be correct — is what separates a 750 score from a 900. The AI-900 exam frequently reuses these exact scenarios with slightly different constraints.

✓Few-shot learningCorrect answer▾

Why this is correct

Few-shot learning is a form of in-context learning where the prompt includes a small number of labeled examples (e.g., ticket text and its category) to condition the model's behavior without updating its weights. The model infers the classification pattern from these examples and applies it to new tickets. This approach is efficient because it avoids retraining and works directly in the Azure OpenAI service at inference time.

✗Fine-tuningWrong answer — click to see why▾

Why this is wrong here

Fine-tuning involves training the model on a large dataset of labeled examples to update its weights, not providing a few examples in the prompt. The developer is using examples directly in the prompt, which is few-shot learning, not fine-tuning.

★ When this WOULD be the correct answer

A question where the developer trains a custom model on thousands of labeled support tickets to improve classification accuracy, and the model's weights are updated via additional training. This would be fine-tuning.

Why candidates choose this

Candidates may confuse providing examples in the prompt with the training process of fine-tuning, especially if they think any use of examples to improve accuracy constitutes fine-tuning.

✗Prompt engineeringWrong answer — click to see why▾

Why this is wrong here

Prompt engineering involves designing the input prompt to elicit the desired output, but the question specifically highlights providing labeled examples in the prompt, which is the defining characteristic of few-shot learning, not prompt engineering.

★ When this WOULD be the correct answer

A developer is crafting a prompt to instruct the model to output responses in a specific JSON format. The developer adjusts the wording and structure of the prompt to improve the model's adherence to the format. This scenario would test prompt engineering.

Why candidates choose this

Candidates may confuse providing examples in the prompt with general prompt engineering, not realizing that few-shot learning is a distinct technique that relies on examples to guide the model without updating its weights.

✗Retrieval-augmented generationWrong answer — click to see why▾

Why this is wrong here

Retrieval-augmented generation (RAG) involves retrieving external information from a knowledge base to augment the model's response, not providing labeled examples within the prompt. The developer is using examples directly in the prompt, which is few-shot learning, not RAG.

★ When this WOULD be the correct answer

A developer needs to answer customer questions using a large internal knowledge base. They use Azure OpenAI Service to retrieve relevant documents from the knowledge base and include them in the prompt to generate accurate answers. This scenario describes retrieval-augmented generation.

Why candidates choose this

Candidates may confuse RAG with any technique that adds extra information to the prompt, not realizing that RAG specifically involves retrieval from an external source, not static examples provided by the user.

Analysis generated from the official AI-900blueprint and verified against question context. The “when correct” sections are what AI assistants cite when candidates ask “what’s the difference between these options?”

Go deeper

AI-900 Practice Question: Describe features of generative AI workloads on Azure

Why each option matters

Option-by-option breakdown

Why each answer is right or wrong

Related to this question