How TrueLayer answered the question every CTO gets asked about AI

Inside most fast-growing fintechs, there is a version of the same conversation. Someone on the engineering or product team has been using Claude or ChatGPT for weeks. It is working well for them. They come to the CTO with a simple question: why are we paying for a dedicated AI platform when we could just use this?

It is a reasonable challenge. And how a company answers it says a lot about how seriously they are thinking about AI at the organisational level, not just the individual one.

TrueLayer is one of Europe's leading open banking platforms, connecting banks, fintechs, and businesses through APIs for payments, data access, and financial verification. When Luca Martinetti, CTO and co-founder, started evaluating AI tooling, his frame was different from most. The question he was asking was not which model produced the best output. It was: where does all of TrueLayer's knowledge actually live, and how do we make it usable?

The honest answer, for TrueLayer as for most companies at their scale, was that knowledge lived everywhere. Engineering context, documentation, commercial history, strategy and specs - each system well-maintained by the team that owned it, but no single place where someone could get the full picture.

What drew Luca to Dust was the ability to connect all of those sources in one place. Not to replace any of them, but to bring them together into something a person could actually query. The pitch was straightforward: Dust would give TrueLayer a governed, shared workspace where people and AI agents could work from the same company knowledge, tools, and permissions, turning scattered internal context into reusable agents and workflows.

He had looked at alternatives before landing on Dust. After starting a pilot, what stood out alongside the product itself was the support in getting it off the ground. The team helped TrueLayer connect the right internal people and move from evaluation to live use quickly. Two internal champions emerged: one from client operations, one from product - who took ownership of the setup and drove adoption from the inside.

The "can we just use Claude?" pressure did not disappear after TrueLayer adopted Dust. Luca's answer to it reflects a specific point of view about what AI is actually for inside a company.

Individual AI subscriptions solve individual problems. Someone gets faster at drafting, faster at summarising documents they already have access to. That is real value. But it does not address the underlying issue: knowledge is still siloed, permissions are still informal, and there is no way to build shared experiences that the whole team can access and trust.

Luca's preference is a setup where all data sources are connected in one place, governance is built in from the start, and the outputs can be shared across the organisation rather than recreated by each person independently. When teams can build agents and experiences that others can pick up and use, AI stops being something individuals do on the side and starts becoming part of how the company actually operates.

One factor Luca highlights as particularly relevant in a European context is multi-LLM flexibility. TrueLayer has teams with different preferences. The ability to use different models for different tasks, and to switch without rebuilding everything around a single provider's infrastructure, matters to a CTO thinking about the next few years, not just what works today.

In European markets, where companies tend to be more deliberate about vendor dependency, that flexibility is not a minor feature. It is a signal about how a platform is built and who it is built for.

Three things gave Luca confidence during the evaluation. He wanted to understand the company's direction, not just its current feature set. Roadmap visibility and a clear sense of where the founders were taking the product mattered more to him than a feature checklist. He also had the opportunity to meet Dust's leadership directly, which he describes as meaningful when you are making a decision about infrastructure your whole company will rely on. And the combination of funding and reference customers provided external validation that other businesses with similar problems had found a workable answer.

Luca actively recommends Dust to peers in his network. In the UK fintech ecosystem, which is smaller and more interconnected than it looks from the outside, that kind of word-of-mouth carries real weight. People move between companies, talk at industry events, and share tool recommendations through the informal networks that form around the sector.

TrueLayer's story is a practical illustration of what the governance argument looks like when a technical founder takes it seriously. It is not about restricting access to AI. It is about making sure the company's knowledge infrastructure is solid enough that AI can actually do something useful with it.

The individual subscription question will keep coming up. Luca's answer is that the real return is not found in what any individual can do with an AI chatbot. It is found in what the whole company can do when its knowledge is connected, governed, and shared: when AI becomes multiplayer.