Vision

AI has moved very quickly. Faster, in some ways, than companies and institutions have been able to change around it.

Models can now read documents, work across large amounts of information, use software, reason through messy inputs and take on work that would have been difficult to automate a few years ago. But when you look inside a lot of important industries, the work still happens through email, spreadsheets, PDFs, old software and people carrying a huge amount of context in their heads.

That gap is interesting to us.

There are now far more problems that software can realistically solve than there were even a few years ago. We do not think the opportunity is to build for all of them. In fact, the abundance of possibilities makes choosing well more important.

Method Assembly is interested in the problems that are still difficult.

The ones where you cannot build a thin product, connect an API and expect the rest to take care of itself. Where you need to understand how an industry works, why the process looks the way it does, what information people trust, where mistakes become expensive and what has already failed before.

Those problems usually take longer to understand. They often involve regulation, legacy systems, fragmented data and people with years of domain knowledge. They are not always the easiest companies to start.

We think that is part of what makes them worth building.

Software is becoming more involved in the work

The last generation of enterprise software was largely built around helping people manage work.

It gave companies systems to store information, coordinate teams, standardise processes and keep records. That created enormous companies, and those systems are not going away.

But AI changes what software can be responsible for.

A system can increasingly read the information itself, investigate an issue, compare options, prepare a recommendation, complete parts of a workflow and bring a person in when judgment is actually required.

That is a different relationship between people and software.

We think many important companies in this generation will be built around that shift. They will not simply add AI to an existing SaaS workflow. They will rethink how the work should happen now that software can take on more of it.

That is the kind of company we want to build.

Why focus matters

There is a temptation right now to build quickly across everything because the technology makes so much possible.

We want to do almost the opposite.

A difficult industry rewards concentration. You have to spend enough time with the problem to notice the things that are not obvious from the outside. The edge cases, incentives, informal workarounds and small decisions that actually determine whether a product becomes useful.

That requires a team to stay with the problem.

So Method Assembly is not designed around producing as many companies as possible. We want to choose a small number of problems where better software can have an unusually large effect, then build focused teams around them.

Each company should be able to stand on its own. Its own users, product, culture and ambition.

What Method Assembly contributes is the method behind starting them: how we research an industry, find the right entry point, build with AI reliably, work with enterprise customers and turn an initial product into something much deeper.

The companies stay focused. What we learn between them compounds.

Where we are starting

We are beginning with procurement and trade finance.

Both are large parts of the economy that still depend heavily on documents, specialist judgment, coordination between organisations and information spread across systems.

They are also difficult areas to build in.

That is intentional.

We want to work on industries where getting the product right matters, where domain understanding becomes an advantage and where a relatively small improvement in how decisions are made can create significant economic value.

There will be other industries after these.

We expect AI to change how work is organised across finance, commerce, logistics, insurance and industrial operations. We do not know exactly where all of that leads yet, and we are suspicious of anyone who claims they do.

What does seem clear is that the underlying capability of software has changed.

That creates room for a new generation of companies.

Method Assembly exists to build some of them.