The key to getting more out of your comms AI prompts

Vague AI prompts might be able to give you a rough first draft, but it won’t know what the audience needs unless the communicator writing the prompt spells it out.

During Ragan’s Communications Boot Camp Virtual Conference last year, Corey Smith, senior manager of employee and change communications at PwC, said that communicators should think beyond the words they’re typing into an AI prompt and instead encouraged them to consider everything a generative AI model might need to provide the best feedback possible.

“Prompt engineering really focuses on the actual instruction that you give the model,” Smith told the audience. “Context engineering is more so everything that the model sees before generating a response. It really helps refine your prompts and your ideas, and ultimately it will produce a better output.”

Smith organizes that information through a framework that asks communicators to define the context and the perspective the AI should take, then specify the deliverable, audience needs and possible risks.