The Enterprise System of Context

How does an organization understand what’s really going on? How does strategy translate into intent, intent into work, and work into outcomes? How do leaders steer something as complex as a modern enterprise?

For most of the SaaS era, the answer to that question was simple: humans.

Humans gathered data in software, organized it into dashboards and workflows, and (hopefully) it guided a person toward a decision. Then *maybe* it recorded the result. CRM, HR systems, ticketing systems, and project management tools all followed the same basic pattern. The human was the operating layer. Software has always been a conduit, but now it’s becoming something much more.

I recently came across Geoffrey Moore’s “Systems of Engagement and the Future of IT” that feels like a time capsule. It’s easy to forget about these past moments when we’re caught up in the current. He captured the previous turn of this cycle well when he described the shift from systems of record to ‘systems of engagement’. Systems of record captured the transactions of the enterprise like orders, employees, financial entries, tickets. Then once those systems were largely built out, the bottleneck moved. The problem wasn’t capturing information anymore. It was coordinating people. Systems of engagement emerged to help organizations communicate and collaborate across increasingly complex supply chains, partner ecosystems, and distributed teams.

He was right about quite a bit. Talking about how we would collaborate, externalize information, and work async. All things we take for granted today. And that logic held for a long time, but now we’re running into the next constraint.

Enterprises are saturated with systems of record and systems of engagement. Chat tools, dashboards, planning systems, whiteboards, docs, ticket queues, meetings, alerts, comments, and now AI agents layered across everything. Communication is abundant, but understanding feels fragmented and misaligned.

The problem isn’t that organizations lack information or communication, it’s that the s*tructure that gives those things meaning* is missing. This problem of organizational context existed before AI. Leaders have always struggled to coordinate teams. But introducing an autonomous agent workforce is bound to amplify this. Over the years we’ve filled the gaps with heroic humans collecting, copying, pasting, and re-contextualizing information. But in this next phase, I believe these roles become what they may have always been, designers of systems.

This is where ‘**enterprise systems of context’** start to matter.

If systems of record captured the business and systems of engagement connected the business, systems of context help the business **understand itself**. They capture the operating model of the organization; the scaffolding that explains how things relate. Strategy connects to initiatives. Initiatives connect to teams. Teams connect to work. Work connects to metrics. Metrics connect to outcomes. Artifacts have properties. Relationships have meaning. The organization becomes legible as a system instead of a pile of disconnected tools and documents.

It’s not a clean hierarchy. We always wanted it to be. We all had the pyramid slide! But it never mirrored reality.

Reality is a messy web of structured, semi-structure, and unstructured data in fragmented systems.

Humans can operate inside ambiguous environments like this surprisingly well. We infer relationships. We recognize patterns. We understand that the same work might be called an epic in one system, a feature in another, and an initiative somewhere else. We carry that context in our heads.

Agents have showed us that this is a superpower of human cooperation. They have a much harder time operating in implicit operating environments.

You can index documents. You can embed conversations. But that doesn’t tell an agent how the organization actually operates. How it’s modeled. And organizational reality at scale isn’t a single graph. It’s a set of overlapping graphs around layers of teams, work, metrics, systems, dependencies. All in constant interaction with each other. Without an explicit model of how those things relate, agents see fragments, not a system.

Systems of context make that structure explicit. They capture the artifacts of the operating model, how they relate, their properties, and what they mean. They turn the amorphous shape of the organization into something machines can reason about.

And once that context exists, something else becomes possible.

Software can begin to **act**.

In the traditional SaaS model, software recorded work after the fact. A salesperson updates the CRM. A manager approves a request. An engineer moves a ticket. The system reflects what happened.

In the emerging model, the system increasingly participates in the work itself. Agents route tasks. They trigger workflows. They escalate risks. They coordinate across tools. They execute outcomes. They even maintain context in the system themselves.

This is the transition toward **systems of action**.

But systems of action can’t operate safely or intelligently without context. To operate autonomously (or even semi-autonomously alongside humans) agents need to understand the structure of the organization they’re acting within.

That’s what systems of context provide. They become the scaffolding that emerging systems of action run on.

Seen this way, the enterprise stack is evolving in a fairly predictable pattern:

- Systems of record captured the facts of the business.
- Systems of engagement connected the people doing the work.
- Systems of context captures, steers, and adapts how the organization actually operates.
- Systems of action execute work within that structure.

The stack doesn’t replace itself. It grows upward as the bottleneck moves.

For the past thirty years, the layer between data and action was a human navigating software.

Increasingly, that layer is becoming structured enterprise context creating explicit operating environments for humans and agent coordination. **H**umans are still very present, but they’re no longer the only thing holding the system together. They’re spending more time designing the system that makes the organization legible.