Why tax season serves as an AI stress test for banks

Hello and welcome to The Buzz, a FinAI News podcast. My name is Whitney McDonald and I'm the editor of FinAI News. At FinAI News, our mission is to lead the conversation on innovation in financial services technology. Joining me today, April 8th, 2026, is Mark Blake, FSI industry practice lead at Stebo Systems.

Mark is here to discuss how tax season serves as a stress test on banking infrastructure, including its AI investment. Thanks for joining us, Mark. Okay. Firstly, thank you for having me on the show, Whitney.

I'm delighted to be here. So I'm Mark Blake. I'm the industry practice lead at FIBO Systems for Financial Services. And we as a company, we specialize in master data management.

What does that mean? It means we help to bring together customer product reference data into that single governed source. So you can look at us as that trusted data foundation that allows banks to power their AI and analytics with that confidence. We bring that regulatory and compliance piece to them as well, Whitney.

Well, the topic at hand today that we're going to go through is how tax season serves as a stress test for AI-driven services implemented at financial institutions. Before we do that, maybe we can kind of set this scene here. We know that financial institutions are investing heavily in AI. That's not a new trend.

That's been the case. We know how much is being invested. But maybe we can kind of just talk through what are those processes where we're seeing AI integrated into financial institutions? What are those standouts?

Yeah, like you've touched on over For the last decade plus, these institutions have been heavily investing in modernization, moving to the cloud, data platforms, and then obviously building on top of that their analytics as well. Now, what I have seen, if I keep it US-specific at the moment, that I think around about two-thirds of their organizations now have in some capacity deployed AI. Some of that may be less strategic than where they want to be in the future. But broadly, that's been, as we know, with the advent of digital assistance, we moved into fraud monitoring, credit underwriting, document classification, and equally as important, risk modeling as well, Whitney.

Now, with AI in place, investment in place, and overall AI strategies implemented, how can this upcoming tax season or the tax season that we're in put stress on those AI systems? systems? Well, obviously, it's a deadline-driven event with the tax pressures that are coming. So from that side obviously it has to be done by that certain point in time That means there peaks So you going to have transaction transaction volumes are gonna spike far higher than what would normally happen due to tax fees and from that side.

So that's obviously the overriding thing here is it's got to be done and completed by a set moment. That brings with it obviously regulatory pressure as well, making sure from that perspective, they're meeting and satisfying the regulators and they need to be confident that their AI models also stand up to that pressure as well. And there's no cracks in that sort of data pipeline and governance. Let's talk through what some of those red flags might be, what some of those gaps might be that could be exposed.

Really, again, reason for companies like Stevo Systems, that gap in data quality. So what you've got here is mismatched records across systems, missing data. All of that helps to undermine the outputs from your AI. So having those inconsistent identifiers for customer as well, they're really some of the major ones.

Behind that, you've also got lineage and auditability issues as well, because, again, you need to be able to explain. explainability now is really paramount to the regulators so how was that decision made can you evidence the decisions that you were made so that you you can get through audit regulatory reviews and then you've got your operational symptoms as well so red flags will show up as you know in manual corrections obviously everything can't be automated at the moment so with that that will bring up other red flags that will impact upon them.

So it could be conflicting information or values in all of those numerous systems that they're looking to access as part of this focus. Now, maybe we can talk here about getting AI ready in terms of your data. Obviously, AI has proven to be that it can be a positive in your institutions. It doesn't just have to be, you know, what cracks are we going to find in the artificial intelligence that has been implemented?

How do you make sure that you have data that can, you know, stand up to a stress test like this? Yeah, and I think what a lot of institutions have learned is that as they've forged ahead, and rightly so, you know, they're embracing AI for the right reasons. I think at the moment, what they're realizing is that from an AI-ready perspective, it's absolutely paramount that you have unified and governed data. So again, what people talk about, that single golden record, it consolidates your customer, your transaction data, and enforces clear validation rules as well.

Again, going back to what I said earlier around explainability and traceability, That really important So each data point has lineage that allows the banks to justify decisions to the regulators and auditors And obviously they want that in real time So real time availability is paramount as well Let's dig in a little bit here. You just mentioned real time. So that's kind of why I want to segue here for a second. But the importance of having access to real time data, especially during tax season, you want access to quick data.

Maybe we can talk about getting that real-time data during this season, especially. How do you ensure that your data is accessible in real-time? And again, this is one of the strengths of your MDM platform. If you have that with something like Stevo Systems MDM, that enables you to have that real-time availability.

So at the moment, a lot of organizations may be faced with manual tasks or they're going in and out of numerous systems. You really want that in a single controlled environment so that, again, from that side, it's going to be less error prone. You can have it in real time rather than obviously taking up huge amounts of those people's time and going off and having to access numerous systems. Now, maybe we can talk here about some lessons learned from tax season or any stressful, you know, stress test or stressful time of the year.

With AI in place specifically, what can banks take away that they can then maybe implement, change or integrate technology around? Yeah, it's obviously one of the key elements of the year. We know we've obviously tax season, but there are others, of course, with regulatory filings that occur through the year as well. But I think certainly from being a focus on the tax season aspect, you know, the AI success depends very much on that data quality and governance.

You must have that foundation in place. And I think that's what a lot of now realizing is that without that, you're going to be faced with the fact that the AI may actually amplify some of those issues that you're faced with if you don't have that foundation in place. Obviously, from an alignment, you know, many of the institutions I touched on have now raced ahead and adopted AI. But really, that's often tended to be more tactical.

and it's maybe siloed so that's not across the whole value chain and it's not being seen by all departments and when you're working on items like this having that true enterprise roadmap is going to be key i think having a coherent enterprise strategy for data and ai governance needs to play into this and then there's the ongoing improvements as well over time you're refine your models, improve processes and personalize. So again that customer experience is in the future getting better and better Maybe we can talk through what some of your clients are approaching you about in terms of you know questions what's coming across your desk, what are financial institution clients really looking for in terms of that good, clean data strategy?

Yeah, we're having numerous conversations with FF institutions, whether that be people directly working in risk, pricing, you know, other areas, credit and such. There's a common theme, though. What they're finding is a lot of the good work they've done so far is maybe not quite yet hitting the mark for them. And what I think they've realised now, and maybe for some they're starting to backfill that, is if you put MDM in place, that really does give you a building block for being able to sort of get what you want from your AI journey and I think the winners in that are going to be those that actually do get those foundations in place and build on top from there really but I think one of the big things for me is ownership it needs to be across the organization then you really start to get the value like I touched on earlier across your own estate your whole value chain by making everyone buy into this and almost you know that whole education awareness ownership understanding if you start to shift the culture and i think that's really important as well then they're going to get the real benefits that they're looking to achieve from their journey on ai yeah i'm glad that you brought that up adoption obviously investment if you if you're investing the kind of money that that these financial institutions have in ai you want it to be used so adoption is definitely key um and then yeah enterprise-wide that's something that we're hearing over and over again um you know ai is not limited to one team or one process or one application um you know it's getting the getting it into the hands of you know entire organizations so absolutely it is across the piece everyone needs to be part of this journey.

And I think those that are seeing that now, who, like you said, have invested heavily, I think ongoing training, upskilling in people is going to be paramount in that. Once you upskill people, that's going to have a big positive for them as well. And I think from that side, again, going back to they're realising now that AI and technology, it's not a technology issue. The one thing that these institutions have is technology.

But if they're going to get that real return on their investment in AI, it's about getting the data sorted. It needs to be fixed. And once it's fixed, then you can build from that. You've been listening to The Buzz, a FinAI News podcast.

Please follow us on Axe and LinkedIn. And as a reminder, you can rate this podcast on your platform of choice. Please be sure to visit us at FinAI News.com for more FinAI News.

Thanks for listening. Thank you.