A blueprint for making organizations legible in the age of AI
Why This Exists
The gap between what an organization knows and what it can act on has always been the central problem of organizational design. AI doesn't fix this gap. It exposes it, loudly, by making the implicit cost of unstructured context suddenly visible and costly. An agent operating inside an organization without an explicit model of how that organization works doesn't know what anything means. It's forced to fill in the gaps and infer at runtime. It acts, sometimes too confidently, on a fragmented map.
At Dotwork, we're a seasoned team of founders, operators, and coaches who have been through many waves of ways of working. We built the leading scaled agile tool acquired by Atlassian in 2019. We've seen every flavor of transformation at Fortune 2000 companies. But this is different. AI will require a paradigm shift.
We don't have all the answers. No one does. But we do know that AI changes the physics and economics of an existing raw material: context. A raw material that has always been important to humans, but it's paramount for agents. Of course, knowledge management has been around, but this transformation requires a different 'wiring' of the organization to support the collection, curation, maintenance and deployment of usable context.
This playbook suggests a blueprint for how to do that. It describes the capabilities required to build an Operating Context System: the infrastructure layer that connects strategy to execution, keeps organizational knowledge alive and current, and gives both humans and AI agents the shared context needed to operate with coherence.
There is no silver bullet. Every organization is different. Treat this as a set of ever-evolving principles in a world that is unrecognizable from quarter to quarter.
This isn't a pitch for any single tool. It's a description of what the system needs to do and why.
A few notes on the words chosen here...
...with the understanding that some of them might be wordy or weird - and are subject to change!
- Operating Context System: 'Context' feels far too general and a 'layer' (we often hear 'context layer') feels too incomplete. A system has structure, ownership, maintenance, and activation.
- Graph-native: Organizational reality is relational. A graph-native system treats relationships as first-class meaning, not as something reconstructed through brittle joins or inferred at runtime.
- Anchor Artifacts: The minimum viable context hubs that 'anchor' the operating context to front-line data. They are the core objects that attract, organize, and refresh relevant context from the tools where work happens.
- Dimensional Artifacts: The durable lenses the organization plans, analyzes, and rolls up around - customers, products, capabilities, markets, value streams, strategic pillars, and similar dimensions.