Payments now do far more than move money between accounts. They influence customer experience, improve operational efficiency and create new opportunities for growth.
Customer expectations continue to evolve as real-time payments, embedded finance and digital experiences become standard. At the same time, ISO 20022 gives financial institutions richer payment data and more consistent messaging, creating opportunities to improve operations and deliver more personalized services.
For many banks, payments have become an important driver of customer satisfaction, liquidity management, fraud prevention and competitive differentiation.
The discussion also explores how open banking and embedded payments are changing customer interactions. Payments increasingly occur within digital experiences, enabling banks to deliver services through trusted partners while creating new opportunities to engage clients.
AI is becoming part of every stage of the payment life cycle. While many organizations first adopted AI for customer support, banks are now applying it to fraud detection, sanctions screening, exception management, reconciliation and payment investigations.
AI also helps institutions use the richer information available through ISO 20022 to improve straight-through processing, reduce operational costs and provide customers with better visibility into payment status.
Rather than replacing payment specialists, AI automates repetitive work and surfaces relevant information more quickly, allowing teams to focus on higher-value activities.
As AI becomes more deeply integrated into payment operations, effective governance becomes essential. Financial institutions need AI decisions to remain transparent, explainable and auditable for regulators and customers alike.
Sean highlights the importance of maintaining human oversight for higher-risk payment decisions, monitoring model performance and preparing contingency plans if AI services become unavailable.
Balancing innovation with resilience helps ensure payment services remain secure and dependable during technology disruptions.
Although payment technology continues to evolve, the institutions that manage a client’s payment flows often build the strongest banking relationships. Understanding payroll, supplier payments, receivables and liquidity positions enables banks to provide more valuable treasury services while strengthening operating deposits.
Andy and Sean also explore how virtual accounts, cash concentration and real-time liquidity management can help banks become trusted advisors to corporate clients.
Looking ahead, emerging forms of digital money, including stablecoins and tokenized assets, will increase the need for unified visibility across traditional bank accounts and digital wallets.
Payments are entering a new phase where speed, intelligence and trust work together. Payment modernization, AI and changing customer expectations are reshaping banking and helping financial institutions build resilient, customer-focused payment operations.
Listen to the full episode to learn how leading banks are turning payments into a source of stronger client relationships, operational efficiency and long-term growth.
- Chapter 1: Why payments have become a business priority for banking executives
-
Andy Schmidt:
Welcome to our latest installment of From Transactions to Trust, a financial services podcast. My name's Andy Schmidt. I'm the global industry lead for banking here at CGI. And I'm joined by my friend and colleague, Sean Devaney, today to talk about payments. Sean, can you take a moment to introduce yourself?
Sean Devaney:
Yeah, hi Andy. Good to speak to you. My name's Sean Devaney. I am part of our product team here at CGI, and I'm responsible for our payments market strategy. I've been doing payments as a job as opposed to just making payments for about the last 35 years or so. So, been involved in most things from physical check processing right the way through to real-time growth settlement systems and real-time payment processing.
Andy Schmidt:
And with that in mind, why are payments becoming a board-level and executive priority for banks right now?
Sean Devaney:
Yeah, it's an interesting question. I think it comes down to a couple of things. One is customer expectations, and the other is that there are just lots of different forms of money now. And that we'll touch on a bit more in the second part of this podcast.
But I think the real reason why it's becoming a strategic opportunity or a strategic discussion for banks is just how much they affect the customer experience now. So, as we move into more and more real-time payments—we've had real-time payments in the UK since 2008 now. But for a number of geographies, that's a relatively new process, certainly for bank-to-bank payments. Seeing how you're moving money not just into your bank, but how you're moving it out, the effect that has on how you view deposits, how you look at managing liquidity, how you undertake fraud detection in real-time processing, how do you look at the data set that you've got behind it, and how do you compete with the other people in your marketplace? And we've seen a real move to standards in this space as well, which I think is generating not just a change, but an opportunity for organizations.
So previously, one of the reasons why it wasn't discussed strategically is because there were so many different mechanisms and methods and standards, et cetera, for making payments. But now we're moving to a much more de facto standard of ISO 2002. I think we're seeing that move driving some behaviors around a consistent set of data for payments, a migration to much richer data sets for banks to be able to use. And then there are things like open banking. So, the open API approach that was started a few years ago and is growing globally now. We're seeing the growth of direct account-to-account payments, not just using for corporates, but for B2C and C2B type payments as well. So that idea of being able to do account-to-account payments brings embedded payments into the conversation. So, it's not just about customers intending to make a payment; it's them going through the normal day-to-day process of their day-to-day activities and making payments without really even having to go onto the bank and make a payment.
So, the best example is the sort of ride-sharing apps where you don't ever touch your wallet; that just comes out of a pre-funded account. That's a really good example of embedded payments, and we're seeing that broaden out. So, I think those things are what's driving it. It's just that ability to have a standardized set of data, which means it's a single conversation about how we're going to do things rather than lots of different conversations for different payment schemes.
It's about the move to real time. So, how do banks deal with that, not just getting money in quicker, but how do they get money out of the organization quicker? And that in turn has an effect on how you manage deposits, how you manage liquidity, how you manage fraud, and how you manage your competitive position in the market.
- Chapter 2: How ISO 20022, real-time payments and open banking are changing competition
-
Andy Schmidt:
That's excellent. And Sean, I mean, there are so many different types of payments you just mentioned. So, you know, how are real-time payments, ISO 20022, account-to-account payments, embedded payments, and open banking changing that competitive landscape?
Sean Devaney:
Yeah, so I think we talked about the different things that are making it more of a strategic board discussion, but each one of those has a different implication, a different impact on how customers are actually using at the payment network. So, the ISO 20022 change makes it much easier for organizations to make strategic decisions about how they're going to deploy payments because now they're able to say, “Well, I know what the message set looks like, I know what my data set looks like, I can deal with that in a much more consistent way across my business.”
I'm seeing a lot more organizations using that ISO 20022 model because ISO 20022 isn't just a payment standard; it's a framework of standards that cover everything from securities through to real-time payments. So that set of data provides a very good basis for organizations to start looking at what their internal data model looks like. And that in turn makes it much easier to share data internally; it makes it much easier to report on things; it makes it much easier to take advantage of some of the richer data that is in those messaging standards.
And then open banking has an impact on how customers interact with their bank. So, the idea behind open banking was always to enable banks to offer services via third parties. So, as a consumer, rather than needing to talk directly to my bank in order to do some transactions, I can use the third party that I want to deal with because it's the person setting up my gym membership or it's the garage that's dealing with my car payment or whatever it might be. I now don't have to deal with my bank to do that. I can go through these third parties.
And that's having a real shift, a real impact on two things. One is how the banks open the data setup that allows that activity to happen. And that goes back to our conversation about the data model and so on. But also, it has an impact on how the banks ensure that they retain wallet share from that customer base. Because now you don't necessarily see the interaction with the customer; you just see the end result of the payment getting made. And so, from a customer point of view, you're interacting with the organization that you want to interact with; you don't necessarily see the bank behind that. So, banks need to see how they differentiate themselves in that market when the customer isn't necessarily using their web app or mobile banking application or whatever it might be.
So, I think we're seeing a lot of organizations looking at this as a way to further sell products through another channel. So, it's not viewed so much as a competitive thing, as in they're not seeing these third parties as competitors, but they're seeing them as opportunities to have another channel through which to sell goods and services. The car finance example is a good one. You could imagine a car dealership being able to offer services where they're pre-validating a customer through to their bank, they're setting up a set of regular payments, they're doing all the credit checks, those credit checks, as well as just using the traditional credit brokers. They can use account information, real-time account information from that customer. You can look at statements and so on. So, you're able to give a much better view of that customer's position, potentially offer them a better product.
The bank has got all that data already because they've got the customer's history, and now the car dealer or the credit broker or whoever it might be can offer a much better deal to the customer. The bank's happy because they're selling a loan through that product, and the customer's happy because their decisions were made more quickly on more up-to-date information.
So, open banking, I think, is really changing that environment. And then that leads in turn, again, onto that whole embedded payments piece. So, you know, that idea of customer journey, a payments journey where you're not necessarily directly interacting with the bank, you're using these services like open banking, et cetera, to facilitate that transaction, whilst you're just connected to the organization that you generally want to connect to, that you're doing your day-to-day activities with.
Andy Schmidt:
Certainly a lot of opportunities there. And you mentioned a way to help the bank not only reposition themselves, but really increase stickiness, increase their wallet share. Other thoughts on how banks can reposition payments from that utility function to a source of growth, to a source of client stickiness and differentiation?
Sean Devaney:
I think a lot of it depends on the type of bank that you're talking to. I think we're seeing quite a lot of organizations looking at where they sit in that life cycle, whether they sit as a traditional transaction bank where they're providing the full end-to-end service to a customer, or whether they're more focused on some of the more niche offerings that they can give to a particular set of or particular segments of customers.
And so, I think we're seeing two things happening. One is large banks that have great connectivity, that have a large customer base, that have a strong correspondent banking network and so on, are focusing on providing that as a service not just to their own customers, but potentially to other organizations as well. White labeling services into markets that those organizations aren't necessarily in, or providing the real-time element of some of those services. We're seeing a lot of growth in that area. But for the right kind of organization that has got the reach and the infrastructure to provide it.
And then on the other end of the spectrum, we're seeing a lot of organizations that don't necessarily want to be in that full transaction banking space because they don't see a way to differentiate by transaction banking. What they see as a way to differentiate is their market segment, their knowledge of the market, their connectivity to particular niche products and services. And then they're using the ISO data sets, the embedded banking, open banking, et cetera, in order to be able to facilitate those types of products that they wouldn't necessarily be able to give otherwise because they don't have access to some of those large data sets that the bigger banks might have.
- Chapter 3: Where AI delivers value across payment operations
-
Andy Schmidt:
That makes sense. Data goes hand in hand with the hot topic right now, AI. And so, when we think about the next theme, AI and payments and looking at how we move from operational efficiency to trusted intelligent payments, where can AI create the most immediate value in payments? Fraud detection, sanction screening, something more mundane like exception handling or recon? Or do you see it going more towards the client-facing elements of it, like liquidity forecasting and client service overall?
Sean Devaney:
Well, the good news is the answer to that question is yes. I think we're seeing it in all of those areas. I think there's a lot of change coming with those AI models. I think at the moment there are three divisions of how AI is being used in our banking clients.
The first is people who are using it to help with servicing a customer through things like help desk complaints and so on, right? So, that's using those AI models to take in large or large-ish data sets, not large language models, but large data sets such as all the regulatory documentation or the internal documentation inside a particular part of the bank or a division or department in the bank, and then using that to be able to help their operators to better serve customers. So, we're seeing a lot of work in that space. And that's kind of the almost table-stakes version of it now, helping those sort of help desk, customer support-type operations.
Then the second one is in things like fraud detection and anomaly detection, et cetera. So, using those AI models in the financial crimes arena to be able to reduce false positives when you're looking at fraud, prioritize alerts when you're looking at particular types of behavior, and also to improve the content that those investigators get when they are actually looking at trying to determine whether a fraud is real or not.
And the third area that I see it being used in is helping to actually make payments on behalf of customers. We're seeing a lot of that happening in the trading space, looking at stocks and shares trading or commodities trading and so on, where people are using those AI models to help manage their portfolio and actually execute trades on their behalf. I think we're going to see more of that happening in the payments market, but I think that the challenge that we've got is actually both a regulatory and liability one.
So, understanding what the impact of these AI models would be on a bank that is making payments that have been initiated by an AI model, I think, is something that needs more work in terms of both the regulations that exist around that, but also for where the liability lies for payments that might be made incorrectly or against a customer's wishes. Either wishes that are beforehand or wishes that come after. Because I think we all know that there can be both of those. So, you can have the intent that you wanted to use AI to be able to do something, but then when actually it comes down to it, then you didn't in fact want it to do that because you lost money.
Andy Schmidt:
Diving a little deeper into that, speed and transparency are always a key issue in the payment world. So how can banks use AI to improve payment speed and transparency without weakening controls?
Sean Devaney:
I think it's important to look at the types of things that AI can do well. And typically, that is looking at very large sets of data and doing it very, very quickly and responding based on a large set of data. So, I think where we can really help without changing the liability model or increasing the risk is looking at things like classifying the reasons for rejected payments. So, you know, what was the reason why the payment got rejected, being able to feed that information back to the customer in real time without having to have somebody investigate and figure out why that payment was rejected.
I think that's quite important.
I mean, fraud things are outside of the scope of that because obviously you've got an element of tipping off that you don't want to do. But for account not found, account closed, those types of rejections, there should be no requirement for somebody to investigate that manually now. That should be coming back through that AI model and getting fed back to the customer. Same for things like status mismatches, where we know that the last time the bank saw the payment was as they sent it out the door. We need to check whether or not the receiving bank got it and sent a response saying that they got it, managing all that using things like a GPI if it's a SWIFT message and so on.
I think using those tools and using AI to manage the data for those tools and feed that back to the client is really key and an area where we can see some real benefit without increasing liability or risk.
We can also look at it for pre-filling in some of the missing data that might be in payments. So, provided that there is somebody there who is checking to see that that data is correct, and it could be the customer that's doing that, we can use AI models to start filling in some of that data so that more payments go straight through once they're sent. And also, just the breaks in the reconciliation process. There are a number of different parties involved in the payment chain; using those AI investigative tools, we can show what was happening to a different payment at different points and show that your outgoing payments and your incoming all match and so on. I think that's an area where AI can really help. And I think that ultimately that helps a client understand where the payment is, why it's delayed, and what's going to happen to it next.
And that's not really so much of a problem when you're making a domestic payment, but anything that you're doing internationally that involves any currency conversion or fees, et cetera, or parties that are outside of your local banking environment, those things become really important. I think that's where AI can really help to provide better services to customers.
- Chapter 4: Why governance and explainability are critical for trusted AI
-
Andy Schmidt:
That makes sense. And I mean, AI, especially with all the data that a payment has already built into it, can be incredibly powerful. And with that in mind, you know, with great power comes great responsibility. So, how should banking executives govern AI and payments to ensure explainability, ensure resilience, ensure regulatory compliance, and of course, as always, maintain, if not enhance, client trust?
Sean Devaney:
You hit the nail on the head, right? That whole governance piece that goes with this is really key. Banks need to have, or be able to show, clear accountability: what made a decision, where that decision was made, and be able to give that information back not only to the customer, but to the regulator as well. So, when the regulator comes knocking on your door, you need to be able to send all of that information back to the regulator to show that you did the right checks at the right places with the right people or the right tools. You've got to be able to explain those decisions that you made, whether that was the person making that decision or it was the AI making that decision.
And that's where some of the challenges have been in the past. Where some of those AI models, if it's a generative model, sometimes you put the same information in twice, and you get two different answers. Being able to reduce the number of times that happens by having more focused, small language models to help make some of these decisions, and by being able to show the logical route the AI model went through to make a decision. It's just like making a decision with a human operator. They would have to justify why they had done a particular thing. The same's true with the AI models.
And then obviously, talking of the humans, then obviously having human oversight for some of those more high-risk cases where you might have a very particularly high value payment being made, you might be making a change to something that changes where the payment goes, the value that it goes for, and so on, they all need some oversight before those payments get made or before that change gets made. And certainly, at this stage in AI use, that's what needs to happen.
And then where we're using it for sort of large-scale things, like we're looking at use of it for fraud detection, for instance, then we need to really keep an eye on the models that we're using to do these fraud detection processes. So that feedback loop of I've made a change to the model, it's had an effect on my straight-through processing rate, it's made it higher, great. But what effect did that have? What knock-on effect did that have on customer complaints? What effect did it have on rejected payments, et cetera? All of those things need to be taken together as an end-to-end story, not just I've changed the model; it improved straight-through processing, for instance.
And also at this stage, I think it's really important to understand what the fallback is when you stop using that AI model. So, not because you intend to stop using it, but because perhaps the regulator has stepped in and said that they don't want you to be using it in that particular circumstance, or because you've decided that actually something has happened and you want to be able to stop that behavior and go back to the way that you did that before. So, it's really important to understand how you do that, what the processes and mechanisms are for going back to that original human-in-the-loop process.
And also, just the AI service might become unavailable. So, we had large outages of a couple of the hyperscalers relatively recently. The same thing can be true for those AI models.
- Chapter 5: How payment flows influence deposits, liquidity management and long-term client relationships
-
Andy Schmidt:
You would think that would also be part of any type of resiliency or BCP type of exercise as well. If the tools we typically use aren't available, what do we do? What's the next step? Because payments need to keep flowing.
Sean Devaney:
Exactly. And that's where we see three different areas where AI could help. When you're just doing it for helping the service desk, that's one thing. If that AI model isn't there, then there are probably manuals that they could use or experienced operators that they can use. When we start making decisions on payments, that just needs to work, right? Or we need to not be using the AI models. We can't say it's okay that only 10% of our payments went to the entirely wrong place. That's fine. That's not the way banks work.
Andy Schmidt:
Right. Being able to keep things moving, even when it's in, shall I say, an incomplete service that's available? That does go a long way towards preserving, not only regulator trust and market trust, but client trust.
Let's shift to our last topic for this podcast: deposit stability and payments. How do you protect that primary client relationship? And with that in mind, how should banks think about the connection between payment flows, operating deposits, and client relationship debt?
Sean Devaney:
That's a really good point, especially when we talk about the idea of certain organizations doubling down and focusing on that transaction banking flow, and then other organizations focusing on perhaps more niche offerings and relying on another organization to do that transaction processing.
I think largely, the bank that processes the core payment flow often is the primary relationship that client has with a bank. They might have some niche providers, and that's great. There's an opportunity for those organizations to grow. But the part of the organization that understands the client the best is the one that understands their payment flows both in and out, understands who they're paying, what for, and what frequency, and so on.
So, I think the organizations that keep control of the payroll processing, their supplier payments to their clients, the accounts receivable, they understand what the overall liquidity needs are for their client and can sweep things between different parts of the organization and then, in turn, interface and feed into the treasury activities of the clients. Those are the flows that create real operating balances and make deposits stable, right?
Andy Schmidt:
How can treasury services, liquidity management, even virtual accounts and cash management influence where clients hold their balances?
Sean Devaney:
I think a lot of it is about having the key capabilities that different organizations do, better or worse than others, right?
So key capabilities will include things like the ability to provide cash concentration. So regardless of what currency that you are operating in, the bank is able to provide you a single view of your position, but also to be able to consolidate into one or a few key currencies.
The idea of having those virtual accounts so that organizations can manage their own internal processes without having to have hundreds of different bank accounts managing different services within their organization. I think that's a really key benefit. We talked about liquidity management and receivables and so on. So, the ability to manage an organization's liquidity properly, the ability to interface that back into treasury systems, CRM systems, et cetera, are really key.
And then we talk a little bit about other forms of money. There are some challenges around things like stablecoins drawing balances away from commercial bank deposits and so on. Actually, I think one of the key things there is not so much whether they draw balances away, because I think we can have a slightly broader discussion about that, but it's about how the bank provides real-time visibility of their balances if some of that balance is on a wallet in a crypto exchange, not in the ledgers of the bank. So, how do they provide that information? How do they provide that consolidated view? I think is key to retaining that sort of stable deposit set.
Andy Schmidt:
Yeah, being that key provider, that key information for them. And as you pointed out rightly earlier, understanding their business and being able to have that sense of timing for when certain transactions should be hitting and if they aren't, what's going on? Are there other things that we need to help with?
Sean, as always, I want to thank you for your time. This brings this installment of From Transactions to Trust, the financial services podcast, to a close. We'll look forward to seeing you on the next episode, where we're going to be talking about digital assets, cross-border payments, and the future-ready payments bank.
Again, my name is Andy Schmidt from CGI and thank you for your time. Have a great day.
