Why Relevant Context Can Still Be Wrong

Semantic similarity is not a policy for use.

Working thesis: A retrieval result can be relevant to the query yet wrong for the action because applicability is multidimensional.

Why this matters for AI agents

An AI agent does not work from organizational reality directly. It works from the instructions, tools, memory, retrieved material and task state that reach its context window. That makes context selection part of the system architecture rather than a cosmetic prompt decision.

For one-off assistance, an imperfect context set may produce an inconvenient answer. For long-running or tool-using agents, the same weakness can persist across steps, be written into memory, propagate to another agent, or influence an external action. The engineering target is therefore not maximum information. It is sufficient, current and applicable information for the task at hand.

This is also why raw retrieval metrics tell only part of the story. A system can retrieve text that is semantically relevant yet still be wrong for the current project, user, environment or point in time. Conversely, an important constraint may have low lexical similarity to the user’s request but still be essential to safe execution.

A concrete example

A coding agent on an EU deployment can retrieve a perfectly relevant US configuration guide or an unapproved newer proposal.

A practical architecture

1. Relevant results. Combine task relevance with scope, freshness, authority and permission checks.
2. Scope filter. Keep the stage contract explicit so inputs, outputs and metadata remain inspectable.
3. Authority / freshness. Apply ownership, scope, lifecycle and approval before this state can be reused.
4. Permission check. Carry identity and permissions through every handoff; visibility is not automatically authorization.
5. Applicable context. Compose the smallest useful task-specific set and retain why each item was selected.

Design principles

• Evaluate applicability separately from relevance.

• Carry task attributes into retrieval.

• Prefer explicit not-applicable signals where explanation matters.

• Keep provenance, lifecycle state and permissions attached as context moves across tools and handoffs.

• Evaluate context quality against the task outcome, not only similarity scores or token counts.

How AuzzurA approaches this

AuzzurA approaches this as a trust and applicability problem for human + AI teams. The useful question is not only what can be retrieved, but what is current, permitted, attributable and relevant to the task.

Questions to ask before implementing this pattern

• What exactly are we persisting: raw source, memory, candidate knowledge or approved knowledge?
• Who owns an item, and who can change its status?
• How is scope represented across organization, team, project, environment, user, agent and task?
• How do we know when an item is stale, superseded or in conflict?
• Can we reconstruct which context reached a participant during a specific run?
• What learning from the run should return to shared context, and what review is required before reuse?

These questions tend to outlast individual model, vector-store and graph-engine choices because they define the organizational semantics around those components.

Sources and further reading

• ContextNest: Verifiable Context Governance
• Governed Shared Memory for Multi-Agent LLM Systems
• Anthropic — Effective context engineering for AI agents