Making Organizational Decisions Accessible to AI

An AI tool can only follow a decision it can find and trust. That means each decision needs a record with what was decided, who decided, when, why, and which earlier decision it replaced, linked to the customers, projects and documents it affects.

• Decisions are made in meetings, calls and chat threads, not in systems of record.
• The outcome is written down, but the reason is not.
• When a decision is reversed, the old version stays in documents and search results.
• Nothing links the decision to the work it changes.

• Decision: the choice, in one sentence.
• Owner: who made it and who approved it.
• Date: when it took effect.
• Reason: the facts and trade-offs behind it.
• Scope: the customers, projects, teams or systems it applies to.
• Status: current, or replaced by a later decision.
• Source: the thread, note or document it came from.

In a context graph, a decision is a node linked to its owner, its sources and everything it affects. When an agent works on a project, it can follow those links to the current decisions and skip the replaced ones. Decisions are the part of organizational memory that most changes how AI should act.

See how East River Labs captures decision history, or talk to us.