To our customers, clients, and investors,
I have spent a great deal of time thinking about what a firm becomes in an economy run on artificial intelligence. This shift is not like the ones that came before it. For decades we used software to make our people faster. What is possible now is different in kind. For the first time, a company can build a real cognitive loop between its people and its systems, where each one teaches the other. That changes the very question we ask about work.
What is at stake is not a tool, or a feature, or which model a company happens to license. It is whether organizations can keep learning, keep building proprietary knowledge, and keep their edge in a world where AI models absorb human and institutional expertise and turn it into a commodity. The risk is quiet, and it is real. Knowledge that took a company decades to earn can be learned by a model in a season.
So every firm now has two kinds of capital to build. The first is human capital: the judgment, the relationships, the creativity, and the pattern recognition of its people. The second is what we call token capital: the AI capability the firm builds and owns. The mistake is to believe one replaces the other. It does not. Human agency is what drives token capital forward. People set the ambitious goals, connect ideas across domains, build the relationships, and notice the patterns that matter most. Without that direction, all you have is compute running in circles.
This is why the real opportunity is not in choosing the best model. Models will keep changing, and the best one this year will not be the best one next year. The opportunity is in the learning loop a firm builds on top of whatever model it uses, the place where human capital and token capital compound together. The firm of the future is the one that compounds learning across its people and its machines, year after year.
That requires a different architecture. Every business should be able to build agentic systems that improve with use while keeping full ownership of the intelligence inside them. The real test of control is simple. A company should be able to swap out a general purpose model and lose nothing of the expertise it has built, the way a veteran employee carries the institution in their head. That expertise has to live in the firm, not in the model it rents.
In practice this means turning your workflows, your domain knowledge, and your hard won judgment into systems that get better every time they run. It means private evaluations that measure whether a model is actually improving at the outcomes your business cares about, not at someone else's public benchmark. It means private environments where models grow stronger on the real work of your organization. It means an institutional memory you can actually query, so nothing the firm has learned is ever lost.
We think of this loop as the new intellectual property of the firm. It is a hill climbing machine, and unlike almost any other asset, it compounds. Every workflow you improve produces better signal, which deepens the tacit knowledge that is unique to you. The companies that begin building this now will hold an advantage that is genuinely hard to copy, no matter what the next model release can do.
We are building toward a particular outcome on purpose. The future none of us should want is one where a handful of models quietly absorb the value of every industry they touch. That world will not hold. There is no public consent for an AI era that hollows out whole sectors. We watched a version of this in the first wave of globalization, where the headline numbers looked healthy while real communities were emptied out, and we are still paying for what it cost. We should not carry that pattern into this era, with a few systems capturing the returns while entire industries find their knowledge commoditized out from under them.
So our priority, and the culture we are building at Charles & Roe, is to help create a frontier ecosystem and not only a frontier model. We want a world where every organization can own the learning loop that encodes its institutional knowledge, and compound its human and token capital on top of it.
This is the ethos we have grown up with. The best platforms enable far more value on top of them than they ever capture inside, and every company should be able to keep innovating and building value of its own. When that holds, companies create value for themselves and for the economy around them. Employees see their expertise amplified, their judgment encoded into systems that make it replicable and scalable, and the benefits flow outward to their companies and their communities.
That is how a firm drives value for itself and for the broader economy. It is the stable equilibrium we should build together, and it is the future we intend to help build.