.png)
Open Banking to Agentic Commerce: What a Decade of UK API Infrastructure Investment Teaches Us
The UK has already run this experiment once.
Roughly a decade ago, the UK's competition regulator required the nine largest current account providers to open standardised APIs to third parties, under the governance of what became Open Banking Limited. The stated goal was competition and consumer choice. The underlying engineering challenge was much harder: get competing banks, fintechs, and regulators to agree on a shared, secure, interoperable standard for machines to move money and data on a customer's instruction.
That is, functionally, the same challenge the industry is now solving again for agentic commerce, except the caller on the other end of the API is no longer always a human clicking a button. Increasingly, it is an AI agent, initiating a payment or a data request on a customer's behalf, within limits the customer has set.
Organisations building or buying into agentic payment infrastructure right now would do well to study how the UK's Open Banking journey actually went, not the sanitised retrospective version, but the real one, with its false starts and its hard-won lessons.
From Human-Initiated API Calls to Agent-Initiated API Calls
Open Banking's core achievement was standardising a specific kind of transaction: a human customer, through a third-party app, authorising a bank to share account data or initiate a payment via a defined, secure API call. The human was always in the loop at the moment of authorisation, even if the underlying data flow was automated.
Agentic commerce extends this model one step further. The agent, not just the app, now initiates the call, operating within constraints the human set in advance rather than approving each individual transaction in real time. The emerging cluster of agentic payment protocols and infrastructure, from card network initiatives to platform-specific commerce protocols, is effectively trying to build the trust and authorisation layer for this shift, much as Open Banking Limited built it for account-to-account payments and data sharing a decade earlier.
This is not a coincidence of timing. It is the same underlying problem, one layer up the stack: how do you let a machine act on a customer's financial life, reliably, securely, and in a way regulators and consumers can trust, without requiring a human to click confirm on every single action?
Lesson One: Interoperability Has to Be Designed In, Not Bolted On
Open Banking's early years were defined by a hard truth: getting nine banks to implement genuinely interoperable APIs to a common standard took longer, and required more governance overhead, than almost anyone initially expected. Divergent interpretations of the same specification, inconsistent sandbox environments, and uneven API reliability across providers were persistent friction points in the standard's early rollout.
The lesson for agentic commerce infrastructure is direct: a protocol that looks standardised on paper is not the same as a protocol that is genuinely interoperable in production, across every merchant, payment network, and agent platform that needs to plug into it. Organisations building agentic payment capability now should assume the same gap will appear, and design integration and monitoring layers accordingly, rather than assuming day-one interoperability across every counterparty.
Lesson Two: Trust Frameworks Matter as Much as the Technical Standard
Open Banking succeeded, where it succeeded, partly because of the technical API standard and partly because of the governance and trust framework wrapped around it: a directory of regulated participants, defined liability and dispute-handling rules, and an accountable governance body in Open Banking Limited that could evolve the standard over time.
Agentic commerce infrastructure faces an even harder version of this problem. When an AI agent, rather than a human, initiates a payment, the questions of liability, consent scope, and dispute resolution become more complex, not less. Who is accountable when an agent misinterprets an instruction and completes an unwanted purchase? What does informed consent look like when the customer set broad parameters days or weeks before the transaction occurred, rather than approving it in the moment?
The organisations most likely to succeed in agentic commerce are the ones treating the trust and governance framework, consent scope, audit trails, dispute mechanisms, as a first-class design problem, not an afterthought bolted onto a working payment protocol.
Lesson Three: Phased Rollout Beats a Big Bang
Open Banking was not delivered as a single flip-the-switch moment. It rolled out in stages: read-only account information access first, payment initiation capability later, and an expanding set of use cases and participants over subsequent years as trust and technical maturity increased.
Agentic commerce infrastructure is following, and should continue to follow, a similar phased logic. The current generation of agentic payment protocols is largely focused on discovery and lower-risk, bounded transactions, extending gradually toward higher-value, higher-autonomy agent-initiated purchasing as trust, tooling, and regulatory clarity mature. Organisations that try to jump straight to full agent autonomy across all transaction types, without the phased trust-building Open Banking demonstrated was necessary, are likely to hit the same friction the earliest, most ambitious Open Banking use cases hit, before the ecosystem around them was ready.
What This Means for UK Commerce and Platform Leaders Today
For a UK-based commerce or platform organisation building agentic payment capability now, the practical implication is not to wait for agentic commerce standards to fully mature before engaging. It is to build on the same principles Open Banking eventually proved out: design for interoperability from the outset, invest in the trust and governance layer as seriously as the technical integration, and roll out agent autonomy in deliberate phases rather than attempting full automation on day one.
Our Agentic Commerce & Payments work applies this directly: building payment orchestration that supports both human-initiated and agent-initiated transactions today, on infrastructure designed to extend as agentic protocols and their surrounding trust frameworks mature, rather than requiring a rebuild each time the ecosystem takes another step forward. For commerce-led organisations already investing in AI-driven buying automation as a route into embedded finance, this Open Banking-informed approach to interoperability and phased trust-building is directly applicable to how that payment orchestration layer should be architected.
The UK's Structural Advantage
The UK is unusually well positioned to lead this next wave, precisely because it has already built the institutional muscle for exactly this kind of standardisation exercise. Open Banking Limited, the Financial Conduct Authority, and the Bank of England have a decade of direct experience governing exactly the kind of interoperability, trust, and phased-rollout challenge that agentic commerce now presents at the payments layer. That is not a guarantee the UK leads the agentic commerce wave the way it led Open Banking, but it is a genuine structural head start that UK commerce and fintech leaders are well placed to build on.
Frequently Asked Questions
What does UK Open Banking have to do with agentic commerce?
Open Banking standardised how third parties access account data and initiate payments on a human customer's authorisation, via secure APIs governed by Open Banking Limited. Agentic commerce extends this same underlying challenge, standardised, trusted, interoperable API access, to a context where an AI agent, not just a human through an app, initiates the transaction within pre-set customer parameters.
What is Open Banking Limited's role in this comparison?
Open Banking Limited is the governance body that developed and maintains the UK's Open Banking standard, including the trust framework, participant directory, and dispute-handling rules that made the technical API standard usable in practice. It is referenced here as a governance model, not as a body that currently regulates agentic commerce protocols.
Why did Open Banking take longer to deliver interoperability than expected?
Getting nine major current account providers to implement a shared API standard consistently required more governance overhead and iteration than the technical specification alone suggested, with early divergence in implementation quality and reliability across providers. This is a documented characteristic of large-scale, multi-party standardisation efforts, not specific to Open Banking alone.
What is the main lesson for agentic payment protocols today?
The three clearest lessons are: design for interoperability across every participant from the outset rather than assuming a written standard guarantees consistent implementation; invest in the trust and governance framework, consent, liability, and dispute handling, as seriously as the technical integration; and roll out agent autonomy in deliberate phases rather than attempting full automation immediately.
Does this mean agentic commerce protocols will follow the exact same timeline as Open Banking?
Not necessarily, and no such timeline claim is made here. The comparison is about structural lessons in interoperability, trust-building, and phased rollout, not a prediction that agentic commerce adoption will match Open Banking's specific multi-year rollout schedule.
How should a UK retailer think about agentic commerce readiness today?
Retailers already investing in payment orchestration and AI-driven buying automation are better positioned to extend that infrastructure toward agent-initiated transactions as protocols mature, rather than treating agentic commerce as a separate future initiative requiring a parallel build.
Conclusion
The UK does not need to guess how a decade-long, multi-party payments standardisation effort plays out. It has already run one, in Open Banking, and the lessons, on interoperability, trust frameworks, and phased rollout, apply directly to the agentic commerce infrastructure being built today.
The organisations that internalise those lessons now, rather than relearning them the hard way, will be the ones ready when agent-initiated payments move from early adoption to default expectation.
Speak With Our Team
Innovify builds payment orchestration and agentic commerce infrastructure informed directly by how UK Open Banking's interoperability and trust framework actually evolved. If you are planning your organisation's agentic payments roadmap, speak with our team to discuss how these lessons apply to your infrastructure.
