The opinions expressed here are those of the authors. They do not necessarily reflect the views or positions of UK Finance or its members.

But artificial intelligence has raised the stakes of that decision exponentially.

A new strategic guide from nCino makes a compelling case for why, when it comes to AI, there’s a better approach.

Unpacking the modernisation challenge behind the legacy “build-first” mindset

The preference for building in-house has traditionally been deeply embedded across much of the UK banking sector. 

It rarely stems from a genuine technical advantage, but rather from legacy system dependencies. Despite a move to integrate third-party AI technologies, these new features are often bolted on to existing infrastructure. The result is a technology estate that consumes vast time and resources just to stay operational and leaves little capacity for genuine innovation.

Cloud-native fintechs have thrived precisely because they carry none of this weight. Unburdened by legacy infrastructure, they deploy new features quickly, respond to market shifts, and in some cases have monetised their technical agility as Banking-as-a-Service offerings. 

No wonder UK banks watching this dynamic are now asking whether the traditional model still makes sense. They’re right to do so because AI has changed the calculus of the debate.

The AI imperative and the new complexity of building intelligence

Building banking-grade AI isn’t simply a matter of hiring data scientists and connecting a large language model. It requires a unified, high-quality data foundation spanning commercial, consumer, and lending operations; domain-specific model training; continuous learning systems; rigorous explainability frameworks for regulators; and an overarching governance infrastructure. 

Each of these is a substantial programme in its own right, but together they represent a multi-year, multi-million-pound commitment, with no guarantee of reaching production. Compounding the issue further is that the talent required- AI engineers, data scientists, and ML specialists- is among the most competed-for and expensive in the technology market.

No wonder.

Building proprietary AI infrastructure is therefore an increasingly unrealistic proposition for UK banking institutions.

A costly race that ends before it starts

UK banks that have tested the build alternative firsthand tell a consistent story. They’ve attempted to develop their own solution before concluding that it simply recreates what a purpose-built platform already offers, but more slowly and at greater cost.

The lesson being that building to match a purpose-built platform is a race that ends before it starts. Taking 5x longer and costing twice as much in the process.

For UK banks weighing AI investment specifically, the gap is wider still. Production-grade AI capability built from scratch takes years, while bought platforms can deliver intelligence from day one, informed by real banking data across hundreds upon hundreds of institutions.

There must be a better way?

The case for a hybrid approach

Build or buy?  Why not the best-of-both?

Today’s dynamic market demands a more sophisticated approach that harnesses the strengths of both strategies while mitigating their inherent limitations and directing scarce development resources for maximum competitive effect.

A hybrid approach whereby banks buy the foundational platforms and infrastructure, then build the institution-specific features and workflows, offers the potential for banks to genuinely differentiate.

This hybrid approach also has direct relevance for UK banks navigating the FCA's expectations around operational resilience, AI explainability, and Consumer Duty compliance. Bought platforms with embedded regulatory frameworks reduce the burden of building and maintaining a compliance infrastructure internally, freeing teams to focus on what matters to clients and the regulator alike.

The immediate impact is better outcomes, faster decisions, and more transparent processes.

Buy the foundation, build the advantage

The banks gaining ground in new markets are not the ones building AI from scratch; instead, they’re the banks that recognised the competitive advantage. The right platform buys a head start on the intelligence that will define banking for a new era.

For UK banks still anchored to the build-first mindset, it’s no longer a question of whether they can build.  The calculus has changed. Today, it’s whether they can afford the time it takes, and more importantly, if by the time they finish, the market will have moved on.

To explore this topic in more detail, download Buy vs. Build: A Strategic Guide for Financial Institutions here.