Definition
Context engineering is the broader practice of assembling and managing the information an AI model needs to produce useful results.
It can include several related techniques:
- Few-shot examples
- Retrieval-augmented generation
- External tool access
- User preferences
- Conversation history
- Context summarization
This differs from traditional prompt engineering in scope. Prompt engineering generally focuses on designing a particular prompt. Context engineering considers the larger system that determines what information reaches the model, how that information changes over time, and which external sources or tools should be included.
Why It Matters
Context engineering helps you think beyond a single prompt — toward the whole system that decides what a model actually sees before it answers.