A governed decision layer for AI.

Give your AI the right story.

Your company has many versions of the same decision. OIDA structures what was recorded, where it applies, what changed and what supports it so AI can work from governed organisational context instead of simply retrieving another relevant document.

Your company doesn't have one clean version of the story.

Decisions live across meetings, documents, threads, policies and tools. Some were approved. Some were replaced. Some only apply to one team, market or project. Some contradict each other. Retrieval can give AI all of them. It cannot, by itself, tell AI which one should guide the next action.

Define the rules. Record the decisions. Resolve what applies.

Organisational ontology

Define how your organisation understands decisions: their scope, type, source authority and governance rules.

Decision registry

Keep decisions together with their evidence, effective dates, history, supersessions and conflicts.

State resolution

When an AI needs organisational context, OIDA evaluates the recorded state and determines what applies.

And when the available evidence is not enough, it says so.

Every answer has a story behind it.

OIDA shows why an answer exists.

NBNorthbeammember · Reviewed governance
Home
Ask OIDA
Knowledge
Decisions
Sources
Ontology
Governance
Review2
Conflicts1
Activity
Scopes
Halden Foods
Norvik Bank
Atlas Mobility
Workspace
Users
Settings
Clients & agents
Sara ContiSara Contisara@northbeam.com
Northbeam/Ask OIDA Halden Foods pitch template 2 clients active today
New conversationNew question
Which template should I use for the Halden Foods pitch?
Sara Conti
Use the 2026 deck template. It applies to every client-facing deck, including the Halden Foods pitch.
SourcesBrand review — meeting notesBrand guidelines 2026 · p. 4
Ask about a decision, a client or a deck… ↵
Answer details
StatusActive
Based on1 decision · 2 sources
Asked bySara ContiSara Conti
ClientHFHalden Foods
Also available to your AI tools
Same answer via MCP

MCP

Use OIDA through MCP. Govern it through the dashboard.

OIDA is not another AI tool to adopt. It exposes organisational state through MCP, so the assistants and agents your teams already use can ask what applies before they act.

~/northbeam — terminal MCP

Privacy

One memory. Each person reads their part.

When someone is invited, they are cleared for the units they work in. The database enforces it on every request, in the dashboard and through MCP: records of other units are not returned, not counted and not used to answer. Company-wide decisions stay visible to everyone, and a question about the whole, asked from a partial view, is never answered as verified.

We publish what we learn.

AI can retrieve every version.
OIDA helps it understand which story to use.

A governed organisational ontology, decision registry and deterministic resolver for AI systems that need more than relevant context.

Talk to us
Demo

Request a demo

Half an hour with the people who built OIDA, on your case: which decision should stay in force, which assistants you use, how the pilot would work.

1 / 7What is your role?
2 / 7What kind of organisation do you work in?
3 / 7How many people work there?
4 / 7Where are you with AI assistants and agents?
5 / 7Which decisions should your AI get right first?
6 / 7When would you like to start?
7 / 7Where should we reply?

Or write to carlo@projectoida.com