A 311 analyst pastes few-shot examples of well-written resolutions before generating a new ticket closeout. What prompt technique is used?

Few-shot prompting teaches style by example. When a 311 analyst pastes well-written resolution closeouts before asking for a new ticket closeout, the model can match municipal tone, structure, and completeness patterns without relying on a bare instruction alone. Those exemplars encode what "good" looks like for residents and supervisors—clear actions taken, next steps, and professional wording—so the new draft inherits that shape.

Zero-shot prompting that drafts a new municipal resolution closeout from instructions directly fails because zero-shot uses instructions alone with no examples, while this scenario pastes few-shot examples of well-written resolutions before generating. One-shot prompting that shows a single municipal resolution example before drafting anew fails because one-shot provides a single example, while the scenario provides multiple well-written resolutions, which establishes the few-shot pattern. Chain-of-thought prompting that shows stepwise municipal reasoning before drafting a new one fails because chain-of-thought is about outlining intermediate reasoning steps, not about showing text exemplars to copy for tone and structure.

Exam caveat: few-shot examples can encode bias or outdated policy if the samples are poor. Curate exemplars that meet current 311 standards; do not treat any past closeout as automatically ideal.

Operational check: maintain a small approved set of anonymized resolution exemplars; paste two to three before generation; have a supervisor spot-check that new closeouts match required fields (action, outcome, follow-up) before the ticket closes.