Persona Prompting: Definition and Examples

Persona Prompting is a fundamental technique in prompt engineering that involves asking a language model to adopt a specific role or identity before responding to a request. By assigning a persona — for example 'You are a senior developer specialized in Python' or 'Act as a physics teacher for high school students' — the user frames the language register, level of detail, and approach angle that the AI will prioritize.

This technique relies on a simple principle: large language models have been trained on texts produced by people with very varied profiles (experts, popularizers, technical writers, etc.). By specifying a persona, one implicitly activates the linguistic patterns and knowledge associated with that type of profile in the training data. The result is a response that is more coherent, better targeted, and often more useful than that obtained with a generic request.

Persona Prompting can be used simply (a single line of context) or sophisticatedly, combining multiple attributes: professional expertise, communication style, target audience, specific constraints, and even personality traits. The more detailed and realistic the persona, the more adapted the response will be to the user's real need.

This approach is particularly effective in professional contexts where a specific expert perspective is needed: marketing writing, code review, simplified legal advice, pedagogy adapted to a given level, or even interview simulation. It often forms the first building block of a more complex prompt system.