Skip to content
Enrich

The language model for banking data

Engage

Personalization and financial wellbeing

 

Drive

Insight exploration and marketing

Assess

Creditworthiness and risk

Company

Everything you need to know about Bud.

Banking on intelligence: The future of banking with agentic models

Before we dive into how banks can become intelligent enterprises, it’s important to qualify the difference between agentic models and agents. Agentic models can perform and complete tasks for us while, when people refer to AI customer agents, they tend to mean LLM-powered assistance.

Agentic models, in my opinion, are a great unlock for us as individuals as they can manage complex finances for us, releasing us from the burden of financial admin (the second largest stressor after health).

I believe that agentic models will very shortly begin to replace labor intensive workloads inside companies. I think this will come as a shock to many industries and, given that banking is one of the most truly data-driven industries, it will change in ways that even neobanks and fintechs may not be prepared for.

Building the banking agent of tomorrow

Imagine a world where banks are no longer burdened by the inefficiencies of human-driven processes. Instead, they’re powered by clusters of agentic models that handle everything from internal workflow, real-time pricing and digital customer ‘segments of one’. Customer operations have already changed forever, see Klarna. This isn’t a far-off, sci-fi concept – it’s the future of banking, and it’s time we start building towards it.

The agentic bank: a new paradigm

Today’s banks are weighed down by bloated ‘mid-office’ operations, creating a huge cost burden that impacts their net interest margin (‘NIM’, but not to be confused with NVIDIA’s AI NIM strategy in this post). 

This financial strain forces many small and medium-size banks to offer uncompetitive products, pushing them into niche markets where larger banks don't care to compete on price. We've seen time and again that better interest rates drive growth, with banks like Kroo, Marcus, and Chase UK leading the charge.

The primary differentiator, in my opinion, is combining smart, driven people with modern technology that results in a ‘more with less’ approach. My view is that agentic banking could be 100x more efficient than a neobank and possibly 1000x a legacy bank, given large incumbents have tens of thousands of middle office, back office and front of house teams. If we break down a bank in its simplest form, it’s a ledger with systems, products, and services, underpinned by a brand. The technology itself doesn't define a bank, but the teams driving innovation and implementing change are the ultimate catalysts behind a bank's performance.

The key insight here is that most of a bank’s decisions are based on numbers and models. So why shouldn’t 60 to 80% of these tasks be handled by intelligent agents? By progressively replacing and augmenting human roles with agents, banks can significantly reduce operational costs, improve efficiency and offer more competitive products. This approach not only transforms how banks operate, but also offers a sustainable model for growth without the need for constant capital infusion.

Targeting the right banks for transformation

The most significant opportunity lies with small to medium-scale banks, particularly those with around 100,000 to 250,000 customers. These could be community banks in the US or building societies in the UK. These types of institutions often struggle with inefficiencies, with fewer skilled employees in their teams and high middle office and IT costs. The idea is to acquire or partner with these banks, transform their operations with agentic models, and turn them into a proof of concept for the future of banking.

A bank of this size offers enough operational pain points to quickly identify areas where AI can be implemented effectively. At the same time, they present a relatively low risk for bold initiatives like shutting down physical branches and reducing human interaction in favor of digital agents. Traditional banks are being challenged by increasing consumer expectations in their digital offerings and a diverse customer population. Agentic models can support cost-optimization and attract a new demographic of customers who decide on their banking partner based on the digital experience offered. By proving the model's efficacy, you can attract a new, more tech-forward demographic and build a new brand on top of the existing one.

Learning from success stories

We’ve seen different models work in the banking sector, such as OakNorth, which ‘industrialized’ the process for SME lending and achieved profitability early on, primarily through a more productive team and streamlined middle office operations. This success provides a blueprint for how agentic models could be used to revolutionize not just lending but all aspects of banking.

A practical implementation: a mass affluent, intelligent bank

One of the biggest gaps in the market is the lack of a mass affluent bank—a bank that positions itself as a ‘private bank’ but is accessible to anyone earning around $60k and above. By leveraging intelligent financial agents, this bank could offer a range of personalized experiences, moving towards a full-time financial assistant model. This would include offering complex financial services like insurance, mortgage and tax services in a highly efficient manner.

Historically, the private banking model has struggled due to the high costs of personal servicing. However, with the integration of agentic models, these costs can be drastically reduced, allowing banks to serve a broader customer base more profitably. The same approach could be applied to customers with high debt-to-income ratios, providing intelligent support for wealth growth.

Conclusion: building towards an intelligent future

The future of banking is not just digital; it’s intelligent. By progressively integrating agentic models into every aspect of banking operations, we can create institutions that are leaner, more competitive and more attuned to the needs of modern consumers. The journey starts with transforming smaller banks that stand to gain great efficiencies, and proving that an agentic model can deliver superior performance. From there, it’s a matter of scaling and redefining the banking experience for the 21st century.

It starts today

Now is the time to start building the bank of tomorrow—one that is driven by intelligence, efficiency and the power of agentic models.

On September 24, we at Bud will begin announcing and demoing some of our first models – so watch this space. It has already been a five-year effort to get to where we are today, but things are advancing faster and faster. I, for one, am bloody excited. 

Find out more about Bud's very first consumer banking agent today