The Invisible Organization Chart

Look at your organization’s chart.

You’ll probably find the usual suspects. Sales. HR. Finance. IT. Operations. Marketing. Clear reporting lines defined responsibilities and accountable leaders.

We’ve spent decades refining how organizations work. We know who owns a process, who makes decisions and who is ultimately accountable for business outcomes.

At least, we used to.

Because something new is quietly appearing in our organizations. Not as a department or a team, but as an invisible layer that sits across them all.

- AI agents.

- Copilots.

- Autonomous workflows.

- Decision engines.

They’re writing proposals, summarizing meetings, reviewing contracts, generating code and assisting employees in ways that were unimaginable just a few years ago.

And yet, they don’t appear anywhere on the organization chart. Perhaps that’s because we’re still treating them as technology. IT owns the platform. Procurement owns the contracts. The business owns the process. Security owns compliance.

But who owns the outcome?

Imagine an AI agent generates 70% of the content for a strategic proposal that wins a multimillion-euro contract. Six months later, that same agent introduces factual inaccuracies following a model update and the quality of proposals starts to decline.

Who notices first? Who is accountable for its performance? Who decides when it is good enough? Who decides when it shouldn’t be used?

These aren’t technical questions. They’re organizational questions.

I sometimes wonder whether we’re asking the wrong questions altogether. We spend a lot of time discussing which models to use, how to build agents and which platforms to standardize on. Those discussions are important, but they’re also the easy ones.

The more difficult question is this: Where does AI fit within our organization? Not technically, but operationally. Tomorrow’s organization chart won’t simply contain people and departments. It will contain capabilities that are partly human, partly automated and increasingly intelligent.

Sales won’t only consist of account managers and proposal specialists. Finance won’t only consist of controllers and analysts. Every department will quietly become a combination of people, automation and intelligence.

The lines between them will become increasingly difficult to draw.

Perhaps that’s why I believe the next challenge isn’t building better AI. It’s designing organizations that know how to operate it.

Because every capability that becomes business critical eventually needs ownership. It needs performance measures, governance, oversight and continuity planning.

We’ve learned how to manage people. We’ve learned how to manage technology. The next decade may require us to learn something entirely different.

How to manage organizations where not every contributor appears on the organization chart.

And perhaps that’s why the most important question isn’t: “How many AI agents do we have?”

It’s simply: “Who is accountable when they don’t deliver?”