Few-shot prompting: Few-shot prompting is including a handful of examples of the input and output you want inside your prompt so the AI model can copy the pattern. It is the fastest way to get consistent formatting, tone, and judgment calls.
Instead of describing your ideal output, you show it. Paste two or three real pairs (here is a raw customer complaint, here is the reply we sent; here is another) and then supply the new input. The model infers the style, structure, length, and even the unstated rules from your examples.
Examples beat descriptions because so much of what you want is hard to put into words. You may not be able to articulate why one product description sounds like your brand and another does not, but the model can pick up the difference from three samples.
Choose examples deliberately. Use real ones, cover the range of cases you expect (including a tricky one), and make sure they are actually good, because the model will faithfully copy your mistakes too. Label them clearly, with something like 'Example 1: Input / Output,' so the model can tell the samples from the task.
Example at work
A sales manager wants call notes converted into CRM entries in a fixed format. She pastes three past call transcripts along with the CRM entries she wrote for them, then adds the new transcript and writes: 'Produce the CRM entry for this call in the same format.' Reps now get identical structure on every call.
Why it matters
Few-shot prompting is the single most useful technique for making AI output match how your team already works. It turns a generic assistant into one that writes your way, and it is often a better first step than fine-tuning or building a custom tool.
Related terms
- Zero-shot promptingZero-shot prompting is asking an AI model to do a task with instructions only, without showing it any examples of the output you want. It works well for common tasks the model has seen many times, such as summarizing or translating.
- Prompt engineeringPrompt engineering is the practice of writing and refining instructions to get reliable, high-quality output from an AI model. It covers giving context, showing examples, specifying format, and iterating on what did not work.
- PromptA prompt is the text (and sometimes files or images) you give an AI model to tell it what you want. A good prompt states the role the AI should play, the task, the relevant context, the output format, and any constraints.
- Fine-tuningFine-tuning is extra training that adjusts an existing AI model's weights using a smaller, targeted set of examples so it performs better on a specific task, style, or domain. It changes the model itself, unlike prompting, which only changes the instructions.
- Chain-of-thought promptingChain-of-thought prompting asks an AI model to work through a problem step by step before giving its final answer. Showing its reasoning tends to improve accuracy on math, logic, and multi-step tasks and makes mistakes easier to spot.