AI ENGINEERING NOW · EPISODE #005
Your AI Agent Doesn't Need More Context. It Needs the Right Context.
Context Engineering: The emerging discipline of designing what AI systems need to know.
11 November 2026 | AI Engineering | Context Engineering
One development. One insight. One action.
1. NOW — What's Changing?
AI engineering is moving beyond the design of prompts toward the broader discipline of context engineering.
Prompt engineering focuses on how we formulate instructions. Context engineering considers the information and operational environment available to the model when it performs a task.
This can include:
- System and developer instructions
- User requests and task requirements
- Retrieved documents and knowledge
- Tool definitions and tool responses
- Conversation history and memory
- Intermediate results from earlier steps
In July 2026, Anthropic described a substantial reduction in the system-prompt material used with newer-generation Claude Code models. The announcement illustrates a broader engineering lesson: more context is not automatically better context.
The central question becomes:
This is not simply a question of reducing tokens. It is a question of relevance, accuracy, timing, structure, and task requirements.
2. WHY IT MATTERS
Modern AI systems can accept large amounts of information. However, a larger context window does not guarantee better results.
An agent may receive:
- Outdated instructions that conflict with current requirements
- Documents unrelated to the task
- Duplicated information
- Tool descriptions it does not need
- Long histories containing irrelevant intermediate work
These additions can increase processing costs and make it harder to identify the information that matters.
Three engineering consequences
1. Reliability: The agent must distinguish authoritative information from irrelevant or conflicting material.
2. Efficiency: Unnecessary context consumes input tokens and can increase processing time and cost.
3. Maintainability: A clearly designed context pipeline is easier to inspect, test, and improve than an unstructured collection of prompts and documents.
Context engineering therefore becomes part of system architecture—not merely prompt writing.
3. ENGINEERING EXAMPLE — A Document-Analysis Agent
Imagine an AI agent responsible for preparing a one-page summary of a client's project risks.
The available information includes:
- 20 meeting transcripts
- 10 project status reports
- 5 contracts
- Several outdated project plans
- The latest approved risk register
Approach A: Give the agent everything
The agent receives every available document at once.
Potential problems include conflicting project dates, duplicated risks, obsolete decisions, and unnecessary context consumption.
Approach B: Engineer the context
The agent initially receives:
- The current project objective
- The latest approved risk register
- The current reporting period
- The required summary format
- Instructions to retrieve additional documents when evidence is missing
The agent can access older documents when a particular question requires them, rather than receiving all historical material automatically.
This example is an illustrative engineering scenario, not a published performance benchmark.
Approach B is a hypothesis to test. If an omitted document contains essential evidence, a smaller context could produce a worse result.
The objective is sufficient, trustworthy context—not simply the smallest possible prompt.
4. TRY THIS — The Context Challenge
Try this small experiment using an AI assistant or an agent you are building.
Step 1 — Choose a task
For example:
“Analyze the supplied project documents and identify the five most important unresolved risks, with supporting evidence.”
Step 2 — Prepare two configurations
| Run A — Full context | Run B — Engineered context |
|---|---|
| All available documents | Current, relevant documents |
| Full background history | Essential task history |
| All available instructions | Current, non-conflicting instructions |
| Same output format | Same output format |
Keep the model, task, and evaluation criteria unchanged. Where possible, use the same model settings and repeat the experiment to check consistency.
Step 3 — Run both configurations
Ask the AI to produce the same deliverable in both cases.
Step 4 — Evaluate the results
Compare each output against the actual source documents and an independently prepared checklist of expected findings.
Do not assume the shorter context will win. Let the evidence decide.
5. SUCCESS SIGNAL — Measure the Difference
Use this simple scorecard:
| Measure | Run A | Run B |
|---|---|---|
| Verified findings (out of 5) | ___ | ___ |
| Unsupported claims | ___ | ___ |
| Input tokens | ___ | ___ |
| Completion time | ___ | ___ |
If token usage is unavailable in your interface, record the approximate context size or use an API that reports token consumption.
Suggested decision rule: Prefer the leaner configuration only when it preserves the required evidence and meets your quality threshold.
For example, if Run B uses fewer tokens but misses a critical risk, the reduction is not a successful engineering improvement.
6. ENGINEERING TAKEAWAY
Context is not just information supplied to AI. It is an engineering resource that must be designed.
An effective AI engineer asks:
- Which information is essential?
- Which information is authoritative?
- Which information can be retrieved later?
- Which information is outdated or unnecessary?
- How will we verify that context changes improve outcomes?
The aim is not to give the agent everything it might possibly need.
The aim is to make the right information available when the agent needs it.
7. GO DEEPER — The Effective AI Engineer
This development connects naturally to an existing case in the The Effective AI Engineer learning series:
CONNECTED CASE STUDY
Case #004 — The Engineer Who Chose the Constraint
Effective engineering is not always about removing limitations. Sometimes the right constraint forces better design decisions.
Context engineering offers a practical application of this idea.
Instead of asking, “How much information can the agent accept?” the engineer asks, “What information is necessary to accomplish this task?”
Learning connection: Use constraints deliberately, then test whether they improve engineering outcomes.
Explore the Effective AI Engineer case-study series and course through AIYING's learning resources.
This connection is an educational interpretation of the case's engineering principle, rather than a claim that the original case was specifically about context engineering.
8. COURSE CHALLENGE — Design a Context Budget
Challenge: Design the context strategy for an AI agent that analyzes project documents.
Prepare a one-page design containing:
- Always include: Essential instructions, current task, and output requirements.
- Retrieve on demand: Supporting documents needed to answer particular questions.
- Exclude by default: Duplicated, obsolete, or unrelated material.
- Verify: A test that detects whether important information has been omitted.
- Measure: Accuracy, unsupported claims, token usage, and completion time.
Suggested deliverable
| Context category | Decision | Reason |
|---|---|---|
| Task instructions | Always include | Defines success |
| Current risk register | Always include | Primary evidence |
| Historical meeting notes | Retrieve as needed | Supports investigation |
| Obsolete project drafts | Exclude by default | May create conflicts |
Completion criterion: Explain why each category belongs in its chosen group, and demonstrate that the resulting agent can still satisfy the task requirements.
Portfolio artifact: A context-budget design note and a two-run evaluation scorecard.
9. Sources and Further Reading
Primary development source
Anthropic, The New Rules of Context Engineering for Claude 5 Generation Models (24 July 2026).
https://claude.com/blog/the-new-rules-of-context-engineering-for-claude-5-generation-models
Technical foundation
Anthropic, Effective Context Engineering for AI Agents (29 September 2025).
https://www.anthropic.com/engineering/effective-context-engineering-for-ai-agents
The document-analysis scenario, comparison exercise, scorecard, and course assignment are original educational applications developed for AIYING. They are not experimental findings reported by Anthropic.
