Open Banking, 1033, and the Agentic AI Catalyst – Steve Boms, Executive Director of FDATA

Open banking in the United States has been on a long and winding road, and the journey is far from over. In this episode, I sit down with Steve Boms, Executive Director of FDATA North America, the trade association representing the fintech companies at the heart of the open banking ecosystem. Steve has been one of the most active voices in shaping U.S. open banking policy for over a decade, and he brings a uniquely informed perspective to where things stand today.

We dig into the current state of the 1033 rule and what amendments are likely coming, FDATA’s firm stance that banks should not be permitted to charge fees for consumer-directed data access, and the growing complexity created by a patchwork of state-level regulations on data privacy, AI, and fintech products. We close with a fascinating discussion on how agentic AI, with its need for clear consent frameworks, robust APIs, and defined liability rules, could become the next major catalyst that finally forces meaningful open banking progress in this country.

In this podcast you will learn:

• The origin story of FDATA in the UK and how it came to the US.
• How Steve has been involved with CFPB and Section 1033 since 2015.
• Over the next 10+ years, how FDATA has been engaged in open banking policy.
• How open banking and open finance has evolved in the UK.
• Who their members are and what FDATA does for them.
• Where we are at today when it comes to the 1033 rule.
• The FDATA view on banks charging fees for access to their data.
• Why this is not really a bank versus fintech fight.
• Why it may be many years before we have a final rule for open banking.
• Why data access negotiations have been put on pause for now.
• What else Steve is working on beyond open banking.
• Why he is increasing concerned about the Balkanization of financial services regulation (see his recent Open Banker column).
• How they coordinate with the other fintech trade associations.
• How they think about the standardization of API and other data standards.
• Why Steve is optimistic about the future of open banking in the U.S.
• Why AI agents could be a catalyzing force for clear open banking rules.

Read a transcription of our conversation below.

FINTECH ONE-ON-ONE PODCAST NO. 572: Steve Boms

Steve Boms

The general view that FDATA has always taken is that this is the customer’s data and the bank should not be permitted to charge fees or monetize that data access in exchange for making it available at a customer’s direction. There are very real, very obvious competitive interests at play when those fees are instituted. In instances where those fees are instituted in ways that differ based on the type of use case that’s accessing the data, you can kind of clearly demarcate where use cases would perhaps dither relative to others and trace that back to where some of the potential revenue streams that the banks have may be implicated. And so our view has always been rooted in section 1033 of the statute that it is at the consumer’s request, meaning free from friction, free from charges, free from any kind of undue burden, that that data must be enabled to access by third parties.

Peter Renton

This is the Fintech One-on-One Podcast, the show for Fintech enthusiasts looking to better understand the leaders shaping Fintech and banking today. My name is Peter Renton and since 2013, I’ve been conducting in-depth interviews with Fintech founders and banking executives.

It’s an understatement to say that open banking in the United States has been on a long and winding journey. And the reality is it is not over yet. My next guest has been an active participant on this journey for more than a decade. Steve Boms is the executive director of FDATA North America, the trade association that represents the fintech companies at the center of the open banking ecosystem.

In this episode, we dig into the current state of 1033 and what amendments you might expect, the contentious question of whether banks should be allowed to charge fees for data access, the growing patchwork of state-level regulations around data privacy, AI, and fintech products. We end with a super interesting discussion about how agentic AI could become the next major catalyst for open banking progress. Now let’s get on with the show.

Welcome to the podcast, Steve.

SB: Great to be here, Peter. Thanks.

PR: My pleasure. So let’s kick it off by giving the listeners a little bit of background about yourself. I mean, you’ve had some interesting stops in your career working for some big names. Why don’t you give us some of the highlights before FDATA.

SB: So I my career on Capitol Hill, working on financial services issues and the Financial Services Committee, went to Capital One, led Capital One’s US Policy Affairs activities, and was then the head of global policy at Yodlee for a little bit of time before heading into FDATA and consulting role.

PR: So you are doing FDATA and you have a consulting business, is that right?

SB: That’s right.

PR: And what’s the, is it sort of a 50-50 split with time? Does it change from week to week?

SB: Changes from hour to hour.

PR: Right, yeah, understood. Maybe before we start talking about FDATA, just tell us a little about your consulting work. Who are you working with there?

SB: We work with financial technology companies, and so they’re all different types of companies and all different types of things. The one thing that they have in common is the existing regulatory structure wasn’t designed with them in mind. Right. And so they’ve had to kind of figure out how to take a 21st century financial technology platform and make it work in a decidedly 20th century landscape. At the same time, there are policymakers in federal and state government who are trying to think about how to modernize existing laws and regulations.

And so we try to help inform those policymakers as to what’s actually happening in the market, how to balance consumer protection with innovation, safety and soundness with technology, all sorts of really complex, interesting and important questions.

PR: Okay, so then now let’s talk about FDATA and you’ve been doing this for a while now. Take us through some of the history and how FDATA came to be.

SB: So FDATA Peter started in the UK back in 2014. And that was at the time when PSD2 was an idea, open banking was an idea on top of that idea, it had not yet been effectuated. And FDATA started as a trade association at the request of HM Treasury, which approached six different fintech CEOs and said, look, we want to get your perspective as we’re kind of designing a system here.

When we need the bank perspective, we know exactly where to go and we get one answer. When we want a FinTech perspective, we don’t know where to go. And when we ask four of you for your perspectives, we get 10 different answers. And so it would really behoove you to kind of coordinate your feedback to us. And the commitment that they made was, if you formalize, we’ll give you a seat at the table, but you’ve got to have a structure in place in order to do that. And so that was the beginning of FDATA as a global organization.

Fast forward to 2017, very beginning of 2018. Several of us who had been active in FDATA in Europe saw a lot of the same kind of trends here in North America. And by that I mean like people were using fintech clearly, right? You had tens and tens and tens of millions of consumers and small businesses using at least one fintech tool. The policymaking audience was looking at it both skeptically, but also in terms of benefit, right?

So on the one hand, there’s all sorts of great additional access and competitive benefits of fintech. On the other hand, as we were saying a little bit earlier, the regulations may not be fit for purpose because they weren’t designed with this in mind. And the third thing was that the market was not solving for this on its own. In the context of open banking in North America, what I predominantly mean is that was right in the middle of the transition away from screen scraping to API access. It was full of friction. It was taking years for data access negotiations to work their way through. And so things generally kind of seemed stuck.

And so we started the North American version of FDATA back at the very beginning of 2018 with the view of trying to figure out how we could create policy guardrails in Canada and the U.S. to facilitate open banking on the other side of the Atlantic.

PR: Interesting. Okay. then given that open banking has certainly evolved differently in the UK versus the US. I mean, remember open banking sort of came with the regulatory mandate, I think it was like 2018 or something like that. It was back a few years ago, but it’s interesting that the adoption here, one could argue is actually stronger with a market led adoption, but you know, it’s been a rocky road, shall we say on the, on the policy front. Maybe you can tell us a little bit about how you were working with, you know, the CFPB section 1033 of Dodd-Frank, which has had a very long and unusual life, shall we say. Tell us a little bit about what the work you were doing back in those days when you started.

SB: Peter, you’ve got to go back into your time machine to kind of go through this story. So obviously, the CFPB gets its authority in 2010 and sits on it for several years. My engagement with the CFPB on this starts in 2015 when the then director of the CFPB goes out into the market at Money 2020 and says, I’m paraphrasing here, we’ve been watching the consumer financial data access marketplace and we have concerns.

We have concerns that incumbents may be restricting access to data, and we have concerns that third parties may not be living up to the right privacy and security standards when they’re accessing that data. And that was kind of the gunshot that signaled the beginning of what has become a decade long and continuing regulatory push in open banking.

So the way that we’ve engaged on the CFPB’s work has ebbed and flowed depending on where they are in their process. It started with, I testified at a field hearing the CFPB did on this in 2017 in Salt Lake City. We’ve provided comments and input into their advance notice of proposed rulemaking. They did an RFI on this also in 2017. And then in 2023, they propose a rule. And so obviously we provided very specific feedback to that rule proposal on things that we thought worked and didn’t. And then they finalized that rule in 2024.

The Trump administration then put out a separate new ANPR on that same rule last year, asking what market stakeholders thought about the existing rule and what amendments should be made. So we commented on that. And then all throughout this 10 plus year history, Peter, I’ve had ongoing communication with the folks in the CFPB who are responsible for this work to do things from just keeping them abreast of market developments and commercial activity to what the rest of the world is doing. Because while we in the US kind of continue to push and pull on the regulation, the world is moving forward in some ways, in better ways, and in other ways, perhaps not as much.

PR: So then are you still in close contact with the folks in the UK and other countries? What do you guys talk about?

SB: Well, we talk about lessons learned, right? What’s worked, what hasn’t worked, what they would do differently if they could do it all over again. So in the UK, for example, Peter, there is this general view of things are not evolving nearly as fast as folks would like from open banking to open finance to open data. The initial thesis of open banking has not proved out the way that folks thought that it was. It has not garnered the kind of payment revolution that everybody thought that it was going to be.

But it has created more competition in other ways in the financial sector. And that’s kind of the rationale for wanting to move more quickly to open finance. In Australia, obviously, a lot of the conversations have been around how to treat screen screening in a post-open banking environment and how you incentivize or disincentivize certain behaviors. So suppose all to say, you know, the benefit of the U.S. being mired right now in the regulatory state that it’s in is there is another opportunity to look at what’s working and what isn’t working and do it differently.

PR: So then FDATA is a member supported organization, right? You’re like a, I mean, do call yourself a trade association? So then why you tell us who your members are? Cause I look and obviously you’ve, know, you’ve got Plaid and MX and Yodlee and some of the open banking platforms that you would expect, but there are others on there. So tell us who your members are and what you actually do for them.

SB: You’ve got about three dozen members, Peter, plus or minus. And so you’re right. Some of the open banking platforms, most of them are members. But then also it’s the entities that interact with that data on the consumer’s or small businesses’ behalf to provide some service or some product or some tool. So I’ll give you some examples. So one example would be Xero, the small business accounting platform.

If you own a small business, you are probably spending much more time than you want managing your books and paying contractors and making payroll. Xero is an online platform that enables you to connect your business accounts, will automatically pull in your transactions so that you don’t have to input every single receipt and track all of your expenses on an ongoing basis. It just makes your life that much easier. And with that information, it can offer you additional services. You can do payroll through Xero, right? You may be able to be connected with an accountant to actually file your taxes when you have to. That’s one example.

Another example would be a company like Betterment, which provides access to investments on behalf of consumers and investor and retirement accounts on behalf of consumers. There again, if you want to fund your Betterment account, because let’s say you want to buy one share of Google, you’ve got to get that money from somewhere. And then you want to sell that one share of Google, you’ve got to be able to put the proceeds of that transaction back somewhere. And so it will connect with your bank account in order to facilitate that transaction.

Rocket Money or Rocket Mortgage. Rocket money is a personal financial management tool. See everything you have in one place, see everything you own in one place. If the first step to financial well-being is make a budget, well, you’ve got to know where you are to do that. So that’s the very kind of…most straightforward example of a fintech tool. And then Rocket Mortgage, like some of our other member companies, uses all types of data to evaluate your credit worthiness as a consumer. So, you know, the old model, Peter, of course, is credit report and credit bureau data, which is helpful, but it doesn’t cover the millions of Americans who are thin file or no file.

But if, for example, you’re applying for a mortgage and you don’t have a very robust credit report, and your lender like Rocket Mortgage said, well, know, Peter, you know, if you would give me access to the last two years of your checking account, I’d love to just kind of see, did you make payments on time? And can we glean anything about your credit worthiness from that information that might actually qualify you for a lower cost product? And so that has enabled access to all sorts of credit products over the years. And many of our members provide those types of credit products, which could be an alternative to higher risk or higher interest loans.

PR: That’s a really good description of the different types of members. So as far as what you’re doing for them, it really you are engaged in Washington, the rulemaking and the, you know, with legislators? I mean, what can you describe what it is that you do on a day to day basis?

SB: Yeah, there’s two parts of it. There is the external facing right to policymakers, both at the state and federal level. And we’re also active in Canada. So we work with the Canadian government, too. And then there’s the internal alignment portion of it as well. So, for example, if you go back to when the Canadian government and the CFPB in the US started their respective processes, they had very specific questions about things like data privacy and liability. And so part of our work is trying to align the fintech community around recommendations or principles on these topics to inform policymakers’ views on them. So what should liability mean in open banking? What type of liability are we talking about? How do you address that liability? The same is true in Canada right now, which is about to enter into a process of crafting an accreditation framework for third parties. And so that’s really kind of technical stuff, right?

How do you accredit a small fintech that two people have started in their garage versus some of the most mature fintechs that have thousands of people and millions and millions of dollars of venture capital behind them? And so a lot of our time is spent behind the scenes trying to come up with alignment within the industry about what those rules of the road should be.

PR: And so then I imagine 1033 has sort of dominated a lot of the work you’ve done over the last few years, but can you maybe just give us a state of play of where we are at today? And I should add that we’re recording this in mid-February because you just never know when some bombshell is going to drop, but tell us where we are.

SB: Yeah, Peter, you might have to give the exact timestamp given how quickly things are changing. But where we are is there was a almost seven year process that the CFPB undertook to craft and then issue a final rule on open banking, the 1033 rule. That rule was published in late October of 2024. There was a presidential election just about two weeks later. And when the Trump administration came into the CFPB in early 2025, they took a very kind of critical look through everything that the previous administration had done. And ultimately, the Trump administration initially sided with litigation that was filed by the Bank Policy Institute, alleging that the 1033 rule almost in its entirety was unlawful. And the BPI litigation effectively argued that the entire rule should be invalidated as a result.

So the Trump-era CFPB first submitted a brief to the court agreeing with the BPI worldview. Over the ensuing six weeks, having heard from the fintech community, consumer advocates, the crypto community, the retail community, effectively most other stakeholders other than very large banks, the Trump administration filed a new brief with the court and said, effectively, we no longer think that the entire rule should be invalidated. Instead, we’d like to craft amendments to the existing rule. And so court, we ask you to basically put this litigation on pause until after we finalize those amendments and the court ultimately agreed with that approach.

So Peter, where we are is we are waiting on those amendments to be proposed. The CFPB, as I mentioned a little bit earlier, put out an advance notice of proposed rulemaking last year, soliciting feedback from the public on what changes they should make. Those were primarily focused on some hot button topics in open banking world, like whether a data provider should be allowed to charge fees and if so, what fees are allowed, whether payment information should be included as covered data under the rule, what the scope of covered banks should and shouldn’t be and covered accounts should and shouldn’t be. And our expectation is that we will see a proposed version of those amendments sometime in the very short future.

PR: Given, I know that Plaid one of your members, they’ve been very famously doing deals with the likes of JP Morgan and others. Does FDATA have an official view on how banks should be compensated for the costs involved in delivering open banking?

SB: We do, Peter. So I’ll say two things. One, I can’t speak for Plaid and any individual company, but in a general sense, it’s important to keep in mind that any company that has agreed to terms that include fees or data access effectively had no other choice because the way that that was structured was one large bank approaching third parties saying, we are going to cut off your access to customer data, even though the customer is directing you to access this data unless you sign this deal. And so you’ve got a binary choice at that point, which is either don’t allow JP Morgan Chase customers in that instance to share access to their data with fintech apps, or try to get the best possible accommodation that you can realizing that all the leverage in the world exists with the bank in question, because there is no substitute for that data, right? It’s not like I’m a JP Morgan customer. It’s not like I can substitute my data with Capital One data or Wells Fargo data. My data lives with JP Morgan Chase in that instance.

The general view that FDATA has always taken is that this is the customer’s data and the bank should not be permitted to charge fees or monetize that data access in exchange for making it available at a customer’s direction. There are very real, very obvious competitive interests at play when those fees are instituted in instances where those fees are instituted in ways that differ based on the type of use case that’s accessing the data. You can kind of demarcate where use cases would perhaps dither relative to others and trace that back to where some of the potential revenue streams that the banks have may be implicated. And so our view has always been rooted in section 1033 of the statute that it is at the consumer’s request, meaning free from friction, free from charges, free from any kind of undue burden that that data must be enabled to access by third parties.

PR: So what you’re saying is that banks have to foot the bill to create the access to the data because it’s their customer that’s wanting their data access.

SB: Effectively, that’s right, Peter, although I would say it bit of a different way. I would say that this is no different than the 21st century version of you being permitted or your bank being obligated to send you a statement by mail. It’s just the digital version of that.

PR: Because there’s a real cost involved and the banks been doing that for hundreds of years, right? Or least decades.

SB: That’s right. And by the way, it’s worth saying, there’s this kind of very common view of open banking is the bank has the data, the aggregator is accessing the data and sharing it with a fintech. The truth, of course, as I know, you know, is much more complex than that. And the data provider can be a bank, but it also can be a fintech. And the data recipient very often is actually a bank, not a fintech. And so this is not really a question. It’s been kind of presented as a bank versus fintech fight. In truth, it’s really a small versus big kind of conversation. It’s about small bank versus big bank, small fintech versus big fintech, and what the economics of this market could and shouldn’t be, and what competition in this market should and shouldn’t be.

PR: Yeah, that’s a really good point because I mean, that’s part of the thing like this, you know, First Community Bank of Bangor, Maine or whatever is not going to be able to negotiate a deal with Plaid. If they try and charge fees, Plaid will say, okay, sorry, we’re not going to play that game. But so just keeping on fees for one last second, you think that the final amendments or addendums to the 1033 rule, what’s your best guess on where they will come down when it comes to fees that banks can charge.

SB: I think the statute is clear. I think the statute says it’s at the request of the consumer, which in other parts of US statute clearly have been interpreted to mean there can be no fee attached. And so I don’t view that the CFPB actually has any wiggle room or room to maneuver here. I think Congress spoke, and absent any amendments to the statute, they cannot allow fees to be charged to recover data.

PR: So assuming that we do know about the Bank Policy Institute’s lawsuit, so the amendments come in and there’s explicitly no fees or they just don’t even mention it. Do we then, I presume, and I know you can’t speak on the behalf of the Bank Policy Institute, but let’s just assume they don’t like that and they, you know, litigate this again. Are we going to be, is it going to be 2030 before we actually have a real open banking working rule that’s effective in this country?

SB: I really hope not, but I’ll acknowledge the outcome that you’ve just described is at least a possible one, right? It’s not at all hard to foresee that that could be. I mean, look, I think if I want to be a little bit more optimistic, what I would say is, you started this conversation, Peter, with the observation, which is entirely correct, that there’s been a tremendous amount of progress on open banking in the US, perhaps more than any other market in the absence of regulatory clarity for going on more than 15 years at this point.

Regulatory clarity or at least direction has been a helpful guidepost for the market over time. I would argue that APIs didn’t really become commonplace in the US until the CFPB first put a stake in the ground and said, we have concerns here. And so just that direction, I think, can be helpful. And so even if ultimately we’re mired in litigation for some longer period of time, a reaffirmation by the CFPB under a different administration that says statute is clear, fees aren’t allowed, would at a minimum provide more clarity to the market of where regulators from both parties and in different administrations have interpreted congressional intent. And so that may be a possible outcome, but I don’t think it would preclude the market from continuing to move forward.

PR: And so when you’re speaking with the people who are your members that are in market, that are open banking data is being used every single day. I mean, no one’s really, I mean, obviously you’ve got the special situation with Plaid and JP Morgan, but people are just moving ahead, right? With their core business and with the rules that we have.

SB: That’s true. But it is also true that the data access negotiations generally have been put on pause for the last better part of a year, while everybody on every side of this market wait for more clarity.

PR: So let’s just move beyond 1033 for a minute because I’d like to kind of get a sense of what else you work on. I’m specifically about data privacy standards. mean, what else do you spend your time on?

SB: Maybe I’ll start Peter high level and then we can get into specifics. So I’ve been thinking a lot recently about kind of the Balkanization of financial services regulation. We’ve got a Republican administration at the federal level in Washington that is overtly on a deregulatory event. Fair enough. But I think it’s a very simplistic view to think that for the market, that just means that it’s an easier path for compliance. Because in practice, what we’ve seen over the last year and even a little bit longer than that is a number of states are becoming much more active on a number of different policy issues that create a much, much, much more complex framework for nationally active companies to do business.

So you mentioned, for example, data privacy. Great example there. A federal data privacy regime that’s preemptive of state law has been the pipe dream of Congress for 20 something years. I’ve been in this space for 22 years. It’s always been talked about, it’s very difficult politically to see how that gets done. In the absence of any kind of federal data privacy framework, you’ve got states like California and Vermont and increasingly New York that are enacting their own laws. Generally, they’re all somewhat in sync with one another, but as time goes on, that’s not going to be true forever.

AI is another great example. You’ve got a number of states, a growing number of states that are either passing AI regulations or legislation in ways that are not always in sync with one another. And so from a financial services perspective, if you’re an AI company, it’s not always clear exactly kind of what you have to do or how you have to do it. In earned wage access and buy now pay later, very, very similar construct. The CFPB initially had rules and interpretive guidance on both of those products. This administration has walked that back. And so now you have a growing number of states that require licensure or compliance with state lending laws on those products. And those are not at all the same across the board.

And so you start to think, Peter, about all of these issues that are very much at the forefront of innovative fintech companies and how they do business. It becomes a very complex dynamic. And so the issue that I spend a lot of my time thinking about is, what are the opportunities to create consumer protective regimes in ways that allow small companies to actually spend their time offering products, services, and tools to consumers and small businesses, rather than paying outside counsel.

PR: So then how do you work with others in the space? You’ve got the American Fintech Council, you’ve got the Financial Technology Association, I mean there are other organizations. How do you kind of interface with these other organizations?

SB: We coordinate closely, especially on issues that we’re all working on together. It’s often the case that some of us are just more granularly focused on some issues and others based on where our membership is and what the issues are that our members are pursuing. For FDATA, we tend to be very, very, very focused on open banking and data privacy. It’s kind of the core reason why we came to be. And so on those issues, we work very closely with them.

I’d also add just it’s also true that there’s a bunch of other stakeholders in the system who care about these things. The merchant community, I would call out, is really, really, really interested in pay-by-bank and other payment rails, both online and in person, as the conversations over debit and credit interchange continue to evolve. So it’s actually a really broad universe of folks who care about these issues.

PR: Where do you come down when it comes to standards, like standardization of APIs and data formats? Is this part of what you think about and work for?

SB: Yes and no. Standards are great. Standards are helpful. Standards are, my oft kind of used metaphor is this is all like electricity in that the average person should never have to think about how this all works. But they want to know is that it’s safe to use, that everything is interoperable. So whether it’s a toaster or a refrigerator or a computer, I just plug it in and it works. And then if something goes wrong, I know who to call and it gets fixed. Standards are the way that that happens.

However, in the absence of clear policy guardrails or clear policy guidance, standards have to make certain assumptions about what the right and wrong choices are for the market. And so until and unless you have clear rules of the road, there are just like natural decisions you have to make to build a standard. What data are we talking about? Right? What accounts are covered? Right access or read access or both? Do we have to have accreditation for third parties who can use our standards?

That is a role in North America, both in Canada and the US, that the government is uniquely qualified to weigh in on rather than standard setting bodies. And so what we won’t do as FDATA is get involved in the actual technical specifications of what a standard should or shouldn’t be. What we try to do is kind of call balls and strikes on what is actually technical build versus what is a policy decision that is a bit outside the remit of pure standard setting.

PR: Okay, so I want to close with just getting your personal opinion again and just as you look at sort of to the future of open banking, I mean, are you optimistic that we’re going to get a framework in place that everybody can live with in the near future? How do you view this playing out?

SB: I am optimistic, Peter. And I’m not just optimistic because I think the statute is clear. I’m also optimistic because if you look at the history of the evolution of this marketplace in the US, it’s always been propelled by inflection points in the market. The next one that’s coming, I think, is agentic AI. And so if you are, let’s say, a large bank, you are going to be very, very thoughtful about what happens when agentic AI tools start accessing your data and potentially making decisions on behalf of your customers. And you want to make sure that your customer is actually given their consent and their informed consent for whatever action that’s going to be.

And so the policy guardrails have to be in place that you know exactly what’s expected of you by your regulators. The technical standards have to exist such that that information can flow safely and securely in each case. And it cannot be facilitated over the long term purely by screen scraping, right?

Although there’s been a lot of movement towards APIs, the overwhelming majority of institutions in the US do not have APIs. And so that data is still generally flowing through screen scraping. So I think the combination of those things is going to propel the industry forward.

PR: That is a really good point. And I know I said it was my last question, but you’ve really, you’ve brought up a thing that, mean, I hadn’t thought about that until you just said it, that the whole permissions kind of system around AI agents accessing data, that is something that needs to be very clear. And it really, seems like it’s a natural outflow of existing section 1033. I don’t mean 1033 did not in any shape or form, I don’t think. perceive that AI agents were going to be accessing data at some point. But clearly, we all know that this is in our near future. Can you just maybe just in closing give us a little bit more meat about how that’s going to work when it comes to permissioned data where we really, it’s the customer’s data, but they want to have an AI agent have access to it.

SB: Peter, it all comes down to what does informed consent really mean and who holds the record of your giving that informed consent. So if, for example, ChatGPT connects to my bank’s API to access my data and then says, with that information, it initiates a $10 transfer out of my account connected to Venmo, for example. What information does Venmo need to make sure that I’m new and authorize that $10 withdrawal. Whose responsibility is it to get that consent? Is it Venmo’s? Is it ChatGPT’s or OpenAI’s? At what point do the AI models need to stop and confirm that you want an action taken from your account that could affect your balance or your account history? What about when it tries to open a loan for you? Right. So if I just type in a prompt into an AI agent and say, like, I’m looking at buying a home, tell me what my mortgage rates are.

What if the AI eventually evolves to a point where it’s actually submitting applications on my behalf because it’s stored my personal information? So Peter, I totally agree with you. It’s a natural outgrowth. It’s a great example of, you know, while policymakers take their deliberative time to kind of think through guardrails, the market is moving much, more quickly. You know, a lot of us were thinking that the next stage in open banking would be write access, not just read access. But I would argue, know, agentic AI might even advance us beyond write access to an entirely new paradigm of how data is shared and what’s done with it once it is shared.

PR: Okay, well that’s, we’ll have to leave it there Steve. That’s really, we can talk another half an hour just about that one topic, but we are out of time. I really appreciate you coming on the show and it was fascinating learning about all the work you’re doing. So thank you.

SB: Thanks Peter.

PR: As I said, I hadn’t thought about this until Steve brought it up, but he makes an excellent point about agentic AI being the next catalyst that forces open banking forward.

Steve’s arguments that AI agents accessing and acting on consumer financial data will demand clear consent frameworks, robust APIs and defined liability rules, it’s a powerful argument. It echoes the history. Screen scraping volume is what finally pushed banks towards APIs. Agentic AI could be the next pressure point that makes the status quo untenable and accelerates regulatory clarity. Much to look forward to there.

Anyway, that’s it for today’s show. you enjoy these episodes, please go ahead and subscribe. Tell a friend, we’ll leave a review. And thanks so much for listening.