Open Banking has been around for several years. With PSD2 in the European Union and Open Banking in the UK, banks were forced to open access to payment accounts and payment services to regulated third parties. Since then, we have seen a growing number of use cases emerge: account aggregation in personal financial management applications, transaction data used for credit scoring and risk decisioning, account verification during onboarding, cashback and loyalty services, and account-to-account payments.
Yet despite all the attention Open Banking has received, its impact remains relatively limited compared with its original promise. Part of the explanation is technical. API quality and availability still vary significantly between banks. Implementations differ, making connectivity complex and costly for third-party providers. Authentication and consent processes can introduce additional friction, while seemingly basic challenges around historical information, joint accounts, transaction references and data quality remain. Aggregators (like Ibanity from Isabel, Digiteal or Tink) have solved part of the connectivity problem, but they also introduce a cost that can make some business models difficult to scale.
More fundamentally, however, Open Banking exposes only a small part of the financial life of a customer. Knowing what happens on a payment account is useful, but it is not enough to understand someone’s complete financial position. Savings, term accounts, investments, loans, insurance and other financial products are equally relevant. And beyond accessing data, many valuable services require the ability to act on it. Open Banking should therefore not be seen as a solution in itself. It is a means to an end.
This is why the evolution from Open Banking towards Open Finance is so important. Initiatives such as the European Financial Data Access framework, or FiDA, aim to extend the principle of customer-controlled data access beyond payment accounts towards a much broader range of financial information. At the same time, schemes such as the SEPA Payment Account Access Scheme, or SPAA, illustrate another important evolution: moving from APIs created mainly to satisfy regulation towards APIs capable of supporting premium commercial services. Dynamic recurring payments, for example, could allow authorised third parties to initiate recurring payments under previously agreed conditions without requiring a new Strong Customer Authentication for every individual transaction.
That shift could fundamentally change how banks look at APIs. Under PSD2, many institutions regarded APIs primarily as a compliance obligation and therefore a cost. In an Open Finance environment, richer APIs can become products in their own right. Better data, higher service levels, additional functionality and premium services can create new revenue streams. The conversation consequently changes from "What are we legally required to expose?" to "Which data and capabilities can we expose to create additional value?". See my blog "Smarter Together: How Data Sharing Will Transform Financial Services" (https://bankloch.blogspot.com/2026/01/smarter-together-how-data-sharing-will.html) for more info.
But why stop at financial services?
Consumers increasingly expect control over their digital assets and information. GDPR established an important principle: personal data belongs to the individual in the sense that individuals have extensive rights over its access and use. Open Data takes this idea further. The real opportunity is to move from data portability to data mobility: not simply "give me a copy of my data", but "allow a service of my choice to use my data, with my permission, when it creates value for me."
Imagine extending this principle beyond banks. Insurers could expose policies and claims. Retailers could expose detailed purchase information. Energy providers could expose consumption. Telecom companies could expose subscription and usage information. Healthcare providers could make health records accessible under strict controls. Governments could expose citizen-related information and services through well-managed APIs instead of attempting to build every digital customer journey themselves.
Combining these datasets changes what is possible. A bank has a remarkably broad, or horizontal, view of a customer. It knows that someone spent €83 at a supermarket, €70 at an energy provider and €45 at a telecom company. The supermarket, however, has a vertical view: it may know exactly which products made up that €83 purchase. Neither organisation has the complete picture. Open Data allows those different perspectives to be combined, with the customer’s consent, to create a much richer understanding of the context in which that customer operates.
This could move digital services away from individual products and towards customer ecosystems and life events. Buying a house is not just about a mortgage. It involves income, savings, insurance, energy, telecommunications, government registrations, utilities and countless retail purchases. Starting a company, moving abroad, having a child or retiring create similarly complex ecosystems. Open Data could allow services to be designed around those customer objectives rather than around the boundaries of individual companies and industries. See my blog "The Next Frontier for Financial Services: Life-Cycle Event Platforms" (https://bankloch.blogspot.com/2025/05/the-next-frontier-for-financial.html) for more info.
Artificial intelligence makes this evolution even more significant. The raw material for increasingly personalised AI services is data. Combining financial transactions, investments, insurance, retail purchases, mobility and energy consumption could enable highly personalised financial assistants capable of continuously identifying opportunities, risks and actions. The competitive advantage would gradually move away from simply owning data towards generating the best insights from data. In an Open Data economy, several organisations may have authorised access to the same underlying information. What differentiates them is how effectively they understand it and turn it into value.
AI agents could take this another step further. Today, exposing an API does not automatically make it useful. Developers still need to discover the API, read its documentation, understand its parameters, build an integration, test it and maintain it. Emerging technologies are starting to change this model. Machine-readable OpenAPI specifications already describe what APIs can do. New standards such as the Model Context Protocol (MCP) allow organisations to expose tools, resources and capabilities in ways AI applications can understand, while agent-to-agent protocols create the possibility for independent AI agents to discover and interact with each other.
The traditional model of application → custom integration → API could therefore gradually evolve towards customer → AI agent → dynamically discovered services and data. Imagine telling a personal financial agent: "Analyse my household expenses and find a way to save €300 per month without materially changing my lifestyle." Instead of being limited to the functionality programmed into a single application, the agent could, with the necessary permissions, discover capabilities offered by banks, insurers, energy providers, telecom operators and retailers. It could analyse expenses, identify alternatives, compare offers and potentially execute authorised actions.
This changes the Open Data question once again. Open Banking asks, "Can I access this account?" Open Finance asks, "Can I access this customer’s financial information and services?" Open Data asks, "Can I access customer-authorised information across industries?" An agentic ecosystem goes one step further: "Can software discover which data and services exist and autonomously orchestrate them to achieve the customer’s objective?"
Paradoxically, this future does not necessarily require organisations to exchange ever-larger quantities of raw personal data. Privacy-enhancing technologies and virtual data clean rooms could allow datasets to be analysed together while the underlying information remains protected. Instead of providing a mortgage provider with years of detailed transactions, for example, a bank could expose a trusted capability that verifies disposable income or affordability. The algorithm accesses what it needs; the external party receives the authorised insight rather than the complete dataset. Open Data does not necessarily have to mean visible data.
This makes trust, consent and data quality critical. Customers will need to understand who can access their data, for what purpose, for how long and what they receive in return. Otherwise, Open Data risks creating a new form of consent fatigue similar to what we already experience with online cookies. And when AI agents start making decisions based on information coming from multiple organisations, semantics and provenance become equally important. An API alone does not tell us whether two companies mean the same thing by "transaction", "merchant", "customer" or "balance". More accessible bad data will not create better intelligence; it will simply allow bad decisions to be made faster.
The road from Open Banking to Open Data will therefore not be simple. Standards, security, identity, consent, API reliability and data quality all need to evolve. But the direction is becoming increasingly clear. In B2B markets, commercial pressure will increasingly force organisations to expose data and services because customers will expect seamless integration. In B2C markets, regulation and consumer expectations are likely to play a stronger role in breaking down existing data silos.
For banks and other data-rich organisations, this requires a fundamental change in perspective. The strategic question is no longer only "Which digital products can we build with the data we own?". It increasingly becomes "Which trusted data, insights and capabilities should we expose so that an entire ecosystem can create services around them?"
Open Banking opened the account. Open Finance is opening the financial relationship. Open Data will open the ecosystem. And AI agents may ultimately become the layer that connects it all together.
Comments
Post a Comment