Company Brain and context
What is context engineering? Definition and examples
In short
Context engineering is the work of giving an AI model exactly the right context for a task: the goal, the rules that apply and the knowledge it needs. It is less about how the prompt is worded than about what the model knows when it answers.
Why it matters: The quality of an AI answer depends more on the context than on the wording of the question.
Prompt engineering
How do I word the question?
Context engineering
What does the model know when answering?
Why does Context Engineering matter for your company?
A model does not know your company. It does not know what was decided last week, which prices apply or which customers come first. Without that context it fills the gap with plausible assumptions.
Context engineering selects, checks and trims what an AI agent needs for a task. Too little context leads to guessing, too much to confusion. For a company this becomes a leadership question: who decides which state applies?
Example: An AI agent drafts the board update
An AI agent is asked to draft the monthly board update. It gets only the prompt “Summarise the month” and access to every document. It finds the old roadmap with “USA 2027” and writes it into the update as the plan.
With context engineering the agent receives the decision “Strategy 2027: focus on the DACH mid-market, no US expansion”, the current figures and last month's update as a template. The draft follows the current state, and the review is about wording, not strategy.
3 common mistakes to avoid
Dumping everything into the context
More is not better. Outdated and contradicting documents distract the model and end up in the answer.
Polishing only the prompt
Better wording does not replace missing knowledge. If the decision is not in the context, the model cannot know it.
Maintaining context per person
When everyone hands over their own notes and prompts, AI agents work from different truths. Shared context needs owners.
Step by step
- 1
Clarify the task
What should exist at the end, and for whom?
- 2
Define the context it needs
Which decisions, rules and sources does this task need, and which not?
- 3
Check the state
Only current, approved content goes to the model.
- 4
Review the results
When an answer misses, first ask: what was missing from the context?
Bring one important decision.
In 30 minutes with Marc or Simon we show where your work drifts from the decision today.
Meet the Company BrainRead on in the PM Lab
Articles that go deeper.
Related terms
- Context window →The context window is the amount of text an AI model can read at once when it answers.
- RAG →RAG (retrieval-augmented generation) is a method where an AI model looks up relevant passages in your documents before it answers.
- Enterprise Context Management →Enterprise Context Management (ECM) keeps a company's context current, reviewed and usable for people and AI agents.