Why AI can unlock Open Finance

Formula 1 cars were always extraordinary feats of engineering, raced by the world’s best drivers. Then in the late 1980s and early 1990s, new technology changed everything.  

Data transmission from the car to the engineers created the ability to make real-time, data-driven decisions on tyre strategy, pit timing, overtaking. Telemetry didn’t make the cars faster in isolation, but it enabled teams to turn every lap into actionable intelligence.  

Since 2018, Open Banking has promised smarter lending and automated money management – so why hasn’t it happened yet? Because we have the car, but not the telemetry.  

In the UK, Open Banking handled 24 billion API calls last year, but turning that data into decisions still needed a team of analysts. The costs just didn’t stack up. But AI changes that – as we move into the Open Finance era, this can be the industry’s telemetry moment.   

The technology will leverage the abundance of rich financial data to make faster and more effective financial decisions. So, what will its impact be?

A fine balancing act

In a recent paper looking at the relationship between AI and Open Banking, the OECD (Organisation for Economic Co-operation and Development) described the relationship as “mutually reinforcing”. Open Finance feeds AI with high-quality, structured financial data, and in turn AI can accelerate the innovation of the last decade to a fundamentally higher level. 

Like any industry, AI’s impact isn’t 100% beneficial. Integrating with Open Finance frameworks comes with what the OECD describes as potential “trade-offs and tensions.”  

For instance, the technology requires vast amount of data, while protection frameworks (such as GDPR) limit how much can be collected. There’s also the issue of consent. Open Finance models are built on customer consent, but AI models develop iteratively, with multiple processing cycles. Getting customer consent at each stage is vital, but will slow down progress. 

The report highlights an existential question for the industry: How can you maintain effective human oversight, accountability and consumer trust and keep pace with the speed of AI’s evolution?  

Where can AI have the biggest impact?

Real-time pricing and idle money

Currently, products are priced based on static credit scores and historical statements. Open Finance and AI would enable banks to analyse live financial fingerprints, with income timings, spending patterns and seasonal cash cycles all influencing the bank’s decisions.  

Using this information, a bank could launch dynamic pricing — giving borrowers with steady income access to lower rates in real-time — or offer embedded products at the moment of need. For instance, if a customer’s rent payment bounces, AI offers a micro-overdraft or short-term loan, removing the need for a payday lender. 

Consumers would also be able to do more with ‘idle money’ sitting in zero-interest accounts. Here’s an example of how this could work: An API sees £50,000 in an account; AI then uses the customer’s spending pattern to identify how much money they need access to as a buffer, then invests the rest across multiple accounts. 

For example: 

• £10,000 in instant access 
• £20,000 in an account with 30-day notice withdrawals 
• £20,000 with 90-day notice required  

AI could also continually monitor the situation and re-optimise how the money is spread as spending and circumstances change.

The OECD report cited the success of this exact use case in Brazil. 1.4 million customers were alerted about dormant funds that could be invested which led to 2 million tailored proposals and generated BRL 14 billion (£2 billion sterling or $2.7 billion USD) worth of transactions in just six months.  

Lending fees and cross-selling 

Currently the underwriting model tends to exclude those with little to no credit history – the self-employed, young people or recent migrants. Just because financial institutions know less about them doesn’t mean they’re a financial risk.  

Using AI to read live transaction data provides more accurate insights: models analyse whether income is volatile or stable rather than relying on a point-in-time snapshot. It means banks can offer loan amounts supported by cash flow evidence rather than approximation, and SME working capital lines can adapt to current circumstances expanding when Open Finance data sees a large client invoice land, contracting once it’s been paid. 

We could also see new payment fee models. The rise of A2A payments puts structural pressure on card and interchange fees, but AI can offset some of the losses. 

If customers pay advisory fees for AI acting on their behalf, banks can get commission by orchestrating the sale of products — their own or a partner’s — earning referral fees or a distribution margin for each new sign-up. 

Finally, AI and Open Finance can make cross-selling more surgical and tailored, rather than being driven by campaign calendars. If a customer receives a large inheritance sum, AI monitoring an account can recognize the deposit and offer immediate guidance. 

Technology and API standards

At the infrastructure level, AI addresses one of Open Finance’s biggest challenges.  Banks with multiple ASPSPs (Account Servicing Payments Service Providers) have to deal with inconsistent data schemas. The new technology can help to normalize data in flight, eliminating manual mapping across every connection.  

ML (machine learning) models could take this further by learning API traffic patterns and pre-emptively scaling or rerouting connections before an ASPSP fails — so banks can act before something breaks, not afterwards.  

On the development side, natural language to API tools would allow developers to describe any changes in plain English; AI then generates the API call, authentication flow and error handling, cutting months of integration work.  

The OECD highlights that AI-driven tools are used to automated repetitive, resource-intensive processes. It gives banks the ideal tool for tracking consent deadlines and renewal channels so data access never lapses, and at checkout, weighing up A2A, card and buy-now-pay-later in real time to pick whatever’s cheapest and best suited to the customer. 

Customer experience

AI can help consumers manage their finances proactively. With a full view of a customer’s financial footprint, AI can detect a shortfall ahead of a direct debit. Money is then moved from savings or a micro-credit line opened to cover it before the customer even notices.  

The same principle applies at the checkout: when making a big-ticket purchase, data confirms affordability and AI offers tailored device insurance, embedded directly in the checkout flow. It could be equally useful for holidaymakers. AI can detect a flight or hotel booking and help with travel money – checking FX exposure and offering to lock in a rate or load a travel card. 

Looking further ahead, Open Finance and FiDA give AI an overview of a customer’s full pension portfolio. It can trigger alerts and corrective action to ensure someone is always on track for their saving goals, decades before retirement. 

For SMEs, the intelligent technology can eliminate the friction between late client payments and urgent invoices that need paying. The shortfall can be covered automatically with a pre-approved credit facility, repaid as soon as the client pays.

Agentic payments

Some of the examples above describe AI that helps to guide decisions. When it comes to scenarios like moving spare cash into savings, covering a shortfall or locking in a travel rate, this is agentic AI. They don’t work without payment initiation, which is why write access is key to unlocking these use cases. It’s why the UK’s sweeping VRP volume grew by 98% year-on-year in 2025 and the success of Canada’s framework depends on its second phase rolling out.  

The use of Agentic AI does raise regulatory questions. Banks will need to understand how it will approach authenticating a non-human making financial decisions and who is liable when something goes wrong. The OECD’s report discusses the challenges presented by agentic agents and maintaining some level of oversight for systems that can “plan, execute and adapt actions with minimal human intervention.” 

AI can be an Open Finance revenue generator

There’s an appetite among consumers for layering AI into their financial lives. But how do banks make sure they’re at the forefront of the industry’s own “telemetry” moment?  

Firstly, start with AI models. The more data they’re trained on, the better they get. So, building your own AI-driven products for use cases like dynamic lending or cash-flow forecasting will accumulate training data that a rival starting three years later won’t have access to. 

Access to this data also relies on customer consent, so invest in managing those relationships, building trust, and ensuring frictionless renewal will be a valuable asset.  

As we’ve discussed in Open Finance and PSD3 articles, it’s shortsighted to view APIs as a compliance cost. Robust, high-performing API infrastructure can become a platform for distributing valuable financial services to customers, fintechs and insurers.  

This opens the door for revenue opportunities such as:

• • Dynamic lending and pricing margins 
  • Replacing interchange with advisory and action-based fee income 
  • Fraud reduction as a measurable profit and loss figure 
  • Fraud-as-a-service for smaller institutions 

Don’t waste the tailwind

The regulations are only travelling in one direction, but each market is arriving at a different pace. The US is rewriting the Section 1033 framework, but the principle of customers owning their own data is at the center of the debate. Canada has announced a two-phase framework that will likely roll out in mid-2027.  

The more mature markets in the UK and Europe are progressing to Open Finance, but via different routes. FiDA is still in trilogue, but once approved, it  will drive the change in the EU and the FCA‘s (Financial Conduct Authority) roadmap will phase in new schemes up to 2030 in the UK.

Soon these capabilities will be expanded from payment accounts across a customer’s entire financial footprint. Banks that build AI monetization infrastructure now will be ahead of the industry when this data becomes accessible. The OECD is quite explicit in the need for financial institutions manage the data effectively: “Many institutions continue to face challenges in integrating external data into legacy systems…(and) without adequate capacity risk either under-utilising available data or deploying AI in ways that are suboptimal or potentially unsafe.”

The core shift that banks need to understand is that you will move from being a balance sheet institution to an intelligence layer. The balance sheet will still be important, but the winners will be institutions that can turn data into the right action at the right moment. AI is the mechanism that will make this economically viable at scale. 

Want to know more about Open Banking?

If Open Banking can turn compliance cost into a revenue stream, introducing AI will scale those opportunities across your customers’ entire financial footprint.

If you want to know more about Open Banking, speak to RedCompass Labs. Our experts have helped banks across Europe get to grips with the latest legislation. Reach out today.