Demand-Driven Context: Teaching AI Agents to Remember What Matters

A framework for enterprise AI agents that stops trying to document everything and starts letting problems drive what gets curated.

Remember Memento? Leonard Shelby — the guy who can't form new memories. Your AI agent is Leonard. Give it a real enterprise problem and it wakes up with expert reasoning but zero memory of your company.

This article introduces Demand-Driven Context (DDC) — a framework where real problems trigger context curation, not top-down documentation mandates. Instead of pre-documenting everything, you let the gaps surface naturally and curate only what matters.

Originally published on Medium.