Why Grolea rewrote itself

For most of the last decade, growth consulting ran on a fairly stable script. You studied a company, found the weak point in its positioning or go-to-market, and handed back a strategy and a plan to run it. Grolea did a version of that for deep-tech and B2B SaaS companies, and for years it worked well.

That script has quietly stopped matching the problem. The founders and executives I talk to do not need another opinion about AI. They have read the same reports I have, and most can describe what good looks like in some detail. What they do not have is anyone to rebuild how the company actually runs now that the tools have changed. That gap, between seeing it clearly and doing something about it, is where almost everyone is stuck, and no strategy deck closes it.

What national scale makes obvious

I get a second view of the same pattern from my work at Sitra, where I lead data and technology initiatives to improve public-sector productivity. The views here are my own. At that scale, with large institutions and serious budgets in play, the technology is almost never what holds things up. The capability is already on the table, available to anyone who reaches for it. The hard part is ownership: deciding what to rebuild, carrying the change through an organization that was never designed for it, and letting go of the old way so the gains can actually land.

None of this is unique to the public sector. A forty-person SaaS company runs into the same wall, on a smaller budget. Everywhere I look, the bottleneck has moved off the technology and onto the harder question of who owns turning it into how the work gets done.

So I rebuilt Grolea

Once that clicked, the old shape of Grolea stopped making sense. Advice, however sharp, does not deliver a rewrite. So I changed what Grolea does. Instead of recommending growth strategies, it now embeds at the executive level, builds an AI-native operating system on the client's own infrastructure, and hands it to the team that will run it. The label I used to lead with, around deep tech, Web3, and strategic advisory, now says more about where I came from than where this is headed.

I will be honest that this was not a tidy repositioning exercise. It was the same rewrite I keep describing, turned on my own practice. If every company now has to rebuild around AI, it would be strange to exempt my own.

Why I am betting on an open stack

The other half of my thinking is out in the open. Paperclip Blueprints is a small tool I built that spins up a complete AI-native company from a short markdown brief. You describe the company in plain language, what it is, what it is for, and what it should not do, and the tool generates the structure an agent organization needs to actually run: the roles and reporting lines, the budgets, the approval gates, the projects, and the operating rules. It builds all of that for Paperclip, an open-source platform that gives a workforce of agents the same operating scaffolding a real company depends on. The tool is early, version 0.1, and it stands on patterns other people worked out before me.

It also carries a conviction I keep getting more sure of as I build: the version of this shift that lasts will run on an open-source AI stack, on infrastructure you control.

The reasoning is plain. If AI becomes the layer your company runs on, the last thing you want is that layer as a rented black box, with your data flowing through someone else's system. An open stack on your own infrastructure keeps the control, the data, and the freedom to change your mind where they belong, which is with you. Building it in the open is how I pressure-test that idea instead of just asserting it.

Early, and worth following

None of this is finished. The pivot is recent, the tools are early, and I am still learning the shape of the work as I go. What I am sure of is the direction, and that the old version of this work is not coming back.

I write about it as I build, roughly every other week, in a newsletter called The Rewrite. If this is the problem you are sitting with, come find it there. It is a better place to follow along than any pitch I could make.