AI0-001 Implementing AI Solutions Practice Question

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 the LLM on API documentation

  Why it's wrong here

  Fine-tuning bakes API documentation into model weights, but it cannot emit structured tool calls or select endpoints dynamically at runtime. Fine-tuning is correct when adapting a model's tone or domain vocabulary, not for invoking external APIs during inference.

• ✗

  Chain-of-thought reasoning

  Why it's wrong here

  Chain-of-thought reasoning structures the model's internal deliberation but produces no mechanism for emitting a callable function name and arguments. It is the right technique for multi-step arithmetic or logic problems, not for binding natural language to API invocation.

• ✗

  Few-shot prompting with examples of API calls

  Why it's wrong here

  Few-shot prompting supplies example input/output pairs, but without a declared tool schema the model cannot emit a structured call the runtime can execute. Few-shot prompting is correct for steering output format or style, not for enabling API invocation.

• ✓

  Function calling

  Why this is correct

  Function calling lets the model emit structured JSON specifying which API to invoke and with which arguments, based on the user's natural-language request. The runtime executes that call and returns results, enabling reliable external API invocation without parsing free-form text.