AI agent glossary
Prompt engineering
The craft of writing instructions and examples that reliably get a language model to produce the output you want.
Prompt engineering covers the choices you make when you write to a model: how you state the task, what context you include, which examples you show, what format you ask for, and what constraints you set.
The most dependable techniques are not tricks. They are the same things that help a capable human colleague: explain the goal and why it matters, be specific about what done looks like, show an example or two, and say what to avoid.
For agents, prompt engineering has grown into context engineering, because the system prompt is only one part of what the model sees. Tool descriptions, retrieved documents, and the running history all shape behavior as much as the prompt itself.
Instead of 'summarize this', a stronger prompt says: summarize this support thread in three bullet points for an engineer, covering the bug, the steps to reproduce it, and what the customer has already tried.
Learn more in the course
Related terms
Go from definitions to doing
Learn to use and build AI agents in 65 practical lessons.