Organizational Flow · In context
People and AI agents in the organization
What changes in architecture and organizational design when AI agents do part of the work — and what does not. A guide to what Organizational Flow, the blog and the Cambium prototype by Christoffer Råsten already say on the subject.
AI amplifies the organization it lands in
The same AI pilot succeeds in one organization and stalls in another, and the technology explains none of it. What decides the outcome is whether ownership is clear, whether capabilities can be named, and whether decisions are made close to the work. Removing technical friction does not remove complexity; it relocates it into ownership, interfaces, information and decisions.
Execution can be delegated; ownership cannot
An agent can execute a decision. It cannot be accountable for having made it. Every outcome needs a named person who can explain it and decide to stop it — which matters more, not less, as agents become more capable.
Read: Owning What AI Can't
Decision rights are read from rules, not switched on
What an agent may do should follow from what the organization has actually settled: act where a signed rule already decided the matter, prepare and stop where judgement remains, ask where no rule reaches. An organization whose agents ask less is one that has decided more things out loud.
Read: Cambium · The Cambium metamodel
The work harness, and why it should be thin
People and agents need the same structure to act with confidence: an outcome, context, boundaries, decision rights, tools and feedback. Its purpose is to reduce the cost of navigating the organization. So the question is not how much context an agent can be given, but how little anyone needs to make a good decision — and AI has made that cost visible for the first time.
When translation stops being scarce
The boundary between IT and the business exists because turning intent into working systems needed scarce specialist competence. As AI makes that translation cheap, business and development move closer together, core IT becomes a platform, and small teams can begin to carry whole capabilities — provided ownership is distributed along with the ability to build.
Read: The IT Department Exists Because Someone Has to Translate · Cheap code gets expensive fast