More Done Right. Not More. Not Less.

More Done Right. Not More. Not Less.

For years, I've seen and felt the tension between what everyone called IT and the rest of the business. It followed a kind of "predictable script". Business moves fast, or at least wants to, IT moves slow. Business finds a workaround, IT calls it shadow IT. IT cleans it up, adds it to the governance backlog, and the cycle repeats.

The implicit answer was always the same: standardize harder. Less tools. Less deviation. Less autonomy. Less is more.

That script has been in my mind broken for a while, but even more so now. AI didn't just accelerate delivery, it made IT fundamentally available to, well anyone. Anyone can spin up a workflow, prototype an integration, or even build a functional tool in an afternoon. The shadow IT problem didn't go away. It got superpowers.

And the instinct, at least in large organizations, is to respond the same way we always have. Lock it down. Standardize harder. Less is more, again.

I think that's the wrong move.

The old tension, briefly

I've spent years inside this problem, and on the cusp looking in. Working with the connectivity layer that makes music streaming, software updates, and vehicle services actually work, the trajectory from engineering decision to customer impact was short and visible. Performance degraded? The customer felt it. Setup was cumbersome? The customer felt that too. Even if they had no idea which service was responsible.

That kind of visibility creates natural discipline. Not because someone mandates it, but because the cost of getting it wrong is obvious, and felt.

Move that same person into an enterprise architecture role, and suddenly the trajectory is much longer. You're advising platform teams, shaping governance models, negotiating tool standards. The customer is still there, but several layers removed. And the temptation, when you can't see the customer directly, is to optimize for what you can see: compliance, cost, control.

That's how you end up defending SharePoint.🙈

The tool trap

Here's a conversation I've had more times than I can count. A team, usually somewhere closer to the customer, shows up with a tool they love. Notion, Airtable, Slite, something modern and well-designed. They've built their entire way of working around it. Team charters, objectives, documentation, all connected. It works beautifully for them.

Then they want to collaborate with someone outside their organization. And it falls apart. Because the other side doesn't have access. Or the decision they made in Notion isn't findable by the people it affects. Or they want to connect it to the corporate AI layer, which means central integration costs that nobody budgeted for.

The principle everyone agrees on, cross-functional collaboration as a core design principle for how the organization evolves, is bulletproof at the principle level. Nobody argues with it. But the moment you move from principle to tool, everything collapses into cost structure, contracts, and fifty thousand people who have spent twenty years working in Excel.

This isn't a tool problem. It's a structural problem. And less is more doesn't solve it.

What AI changes

The reason this matters now more than ever: AI removes the last remaining argument for slowness.

The old defense of IT governance was partly legitimate. Standardization reduces security risk. Common tools reduce training costs. Shared platforms enable auditability. These things are real. And when the alternative was "wait six months for IT to procure something," the friction was at least somewhat manageable.

Now the friction is gone on the other side. You don't wait six months anymore. You build it this afternoon. Which means the gap between what the organization mandates and what people actually use is going to widen, fast, unless the governance model evolves with it.

The wrong response is to double down on less. You won't win that fight. Not anymore.

But "more is more" is equally wrong. More tools, more integrations, more autonomy without structure doesn't give you speed. It gives you chaos with a better UI.

More done right

What I'm landing on, both as a practical stance and as a principle, is this: the goal isn't to reduce proliferation. It's to raise the floor.

You can use whatever tools you want inside your organization. But if you take a decision that affects someone else, it goes somewhere accessible. If your tool incurs central cost, you fund it. If you want to collaborate cross-functionally, you use the common layer, not because it's the best tool, but because it's the one everyone can reach.

That's not less is more. It's not more is more either. It's more done right. Proliferation is fine. Chaos is not. The job is to define what "done right" means clearly enough that the boundary is obvious, and then hold it.

The interesting shift with AI is that "done right" now has to include things we didn't used to worry about. Who owns this workflow? What decisions did this agent make, and where are they documented? If something breaks across three AI-assisted processes, who explains it?

These aren't IT governance questions. They're organizational maturity questions. And they don't get easier with less tooling. They get easier when the principles are clear and the structure holds.

The part nobody wants to say

Most large organizations won't get there. Not because the principle is wrong, but because naming the problem clearly requires admitting that years of "less is more" governance created the conditions for shadow IT in the first place. That the cost structure of Microsoft contracts and Excel-ingrained workflows is partly a self-inflicted constraint. That the teams using Notion aren't going rogue, they're compensating for something the center failed to provide.

That's a hard conversation. But it's the right one. And AI just made it urgent.

/M