Context engineering is the practice of deciding what an AI model should see before it generates an answer.
Teams often focus on prompts or model choice before they fix the context problem. Context engineering forces a simpler question: what should the model know for this turn, and why?
That includes working state, retrieved documents, and persistent memory. Good systems treat those as distinct layers rather than one giant prompt blob.
It is scoped, selective, and explainable. The system can say why a piece of context was added and which boundary it belongs to: project, user, or session.
That is what makes context engineering a product and infrastructure discipline, not just a prompt-writing trick.
Use these glossary pages and commercial landing pages to move from definition to implementation.