Prepare and test your skills
Prepare and test your skills
Free sample question, answered and explained below. The Generative AI Leader practice bank has 2,191 more like it. Try 10 free: instant score, no account.
A customer service team wants to build an automated workflow with a large language model. The process requires several sequential actions:
When attempting to execute all tasks in a single large prompt, the model frequently misses specific policy rules and produces inconsistent outputs.
Which prompting strategy should the team adopt to decompose this complex workflow into reliable, interdependent steps?
Few-shot prompting provides the language model with one or more explicit input-output examples (shots) within the prompt context to demonstrate the desired format, style, or reasoning pattern before asking it to handle a new query.
Few-shot demonstrations are frequently utilized within individual links of a prompt chain (such as formatting the JSON intermediate step), but few-shot prompting itself is not the architectural pattern for task decomposition.
Zero-shot prompting involves giving a model a task instruction without providing any demonstration examples or intermediate operational steps. The model must rely solely on its pre-trained knowledge and the immediate instructions to generate a response.
While individual prompts within a chain can utilize zero-shot prompting, zero-shot alone is not the architectural pattern that coordinates sequential, multi-stage workflows.
Prompt chaining is an advanced prompt engineering technique that breaks down a complex, multi-stage task into a sequence of smaller, discrete prompts. In a prompt chain, the output generated by one prompt serves directly as the context or input for the subsequent prompt in the workflow.

Attempting to perform extraction, structured formatting, domain verification, and creative text generation in a single prompt introduces cognitive overload for the language model, leading to missed instructions and hallucinations. Prompt chaining directly resolves this by creating deterministic, verifiable handoffs between focused model invocations.
Role prompting (also known as persona prompting) instructs the generative model to adopt a specific identity, perspective, or professional persona (such as "Act as an experienced customer care specialist" or "Act as a compliance auditor") during response generation.
Role prompting can be incorporated into the final drafting stage of a prompt chain, but on its own, it cannot decompose complex tasks into interdependent steps.
Answer them yourself and get your score the moment you finish. No account needed.
Full access includes
Keep the momentum going with these hand-picked practice scenarios
Last updated