Exams / AIF-C01 / Generative AI concepts / Question 4
AIF-C01 practice question: Generative AI concepts (4 of 5)
Fundamentals of Generative AI A free question from our original AIF-C01 practice exam. Try it, then open the answer to see why each option is right or wrong.
An insurer's claims assistant gives vague answers. The prompt wording is already clear, but the model never sees the customer's policy, the conversation so far, or the tools it could use to look up a claim. The team decides to build a system that assembles exactly this information into the model's context window for each request. What is this practice called?
Choose one answer.
- AA. Pre-training, because the model learns the policies from scratch
- BB. Context engineering, because it decides what information the model is shown
- CC. Prompt caching, because it stores the conversation for reuse
- DD. Hyperparameter tuning, because it adjusts the model to the insurer's data
Show the answer and explanation
Answer: B
- A. Wrong: Pre-training builds a model from a very large dataset. It does not supply per-request information such as a customer's policy.
- B. Right: Correct. Context engineering dynamically assembles the right instructions, retrieved data, history and tools for each request.
- C. Wrong: Prompt caching reuses an unchanged prompt prefix to cut cost and latency. It does not choose which information to include.
- D. Wrong: Hyperparameter tuning optimises training settings. It does not change what the model is shown at inference time.
Explanation
Prompt engineering focuses on what to ask. Context engineering focuses on what to show: it designs the system that fills the context window with the system prompt, retrieved documents (RAG), conversation history and memory, the user's profile and tool definitions (for example from MCP servers). The quality of the output is limited by the quality of that input.
Study
Keep practising
Answer questions with a score at the end, or practise just this topic with the explanation after each one. No sign-up needed.