Everyone seems to think AI will decimate the people who are closest to the problem, I don't.
While the world has been racing to announce the next model, the next benchmark, the next trillion-parameter breakthrough, we have been busy. Busy believing something that sounds almost naive in 2026: humans will always create better outcomes the closer they are to the problem.
Not closer to technology. Closer to the problem. There is a difference, and the firms that are the closest partners have understood it for decades. You sit in the room with the CFO who cannot sleep. You see the supply chain that breaks every quarter. You know which process actually matters and which one is theater. That proximity, that knowledge, is not something an LLM can replicate. It is the most valuable asset in enterprise AI, and it already sits inside your firm.
The industry is telling your clients they don't need consulting anymore. That agents will replace advisors. That software will eat consulting. We think that is exactly backwards. The harder the AI problem, the more it needs a human who understands the business context, the stakeholders, the politics, the compliance constraints, the thing that make this client different from every other. Trust is not automatable. And that is the difference between your firms and what the hyperscalers sell. LLMs are a commodity, but to bring them into production you want reliability, ownership, and to be sure of IP and data rights.
AI is going to eat software, not people. But most of the VCs still coming down from the decade of building cookie-cutter SaaS have a hard time seeing it yet, and so does Silicon Valley. I think one of the most prescient memos of this new age is "T here Are No Moats. " I just don't think most people understood what it meant. It extends far beyond LLMs and foundation models. Everyone will make software, everyone can build something, so software won't matter. We spent decades turning humans into robots in front of SaaS applications just to sell more seats, but now anyone can do the opposite, build what matters to them. That means there will be a Cambrian explosion of software. Outside the only moats that remain, this will be chaos, and it will make it almost impossible to sell without trust. We will see marketplaces, or feeds (TikTok for software), the 15 seconds of app fame.
My belief is that this will strengthen the role of those whom companies trust tremendously, and let's be real, no one really trusts software companies or Sam Altman very far. Where are the returns from the last decade of investing in more? AI is flipping everything on its head. It's the most human interface ever made, but for it to work, software has to disappear. That's the tension in the market today.
We have taken a different route for the last few years, believing that the only place that has moats big enough to allow us to build something differentiated enough is in a market where no one can really tell Lovable, Replit, or the other tools apart. We certainly made lovable software, but more than anything we aimed to make software disappear and make AI workable in the hardest environments on earth: healthcare, compliance, payments, the Fortune 500. That is not an LLM problem, that is a trust-at-scale problem. That is why we see consultants and their partners all over the world working with businesses every day to help them.
So we built a platform around a single concept: the workblock. A workblock packages human expertise, AI, and business logic into something both static and dynamic. Static in that it is reusable, governed, auditable, a permanent asset your firm owns. Dynamic in that it runs against live systems, learns, writes directly into ERPs, processes millions of real transactions. Consultants build them. Our clients run them. All interoperable from day one, create, link, make workflows, AI embedded natively in every process.
Most software companies charge clients a license and then compete with the clients partners for the client relationship. We do the opposite. We try to empower the humans who spent decades learning about a single company's processes, to remove the middleman, and let every consultant in the world leave something more behind than just a PowerPoint or a spreadsheet. Instant digital transformation, without the overpromising of software vendors to destroy brands and trust when it doesn't work. Best of all, we don't do licenses. We only ever charge for outcomes. That way we are aligned with our partners and clients from day one.
Boutique, global, specialist, generalist, we think all types of consultancies matter. What matters is that you believe what we believe: the humans closest to the problem will always outperform the ones furthest from it, no matter how good the model gets.
First published on LinkedIn, 12 March 2026.