• Artificial intelligence (AI) needs to work with a solution for it to provide great results
• Without knowing the benefits, users will see AI as a threat rather than a compliment
• In terms of real-time payments, if something goes wrong, SWIFT’s stop and recall function for gpi will help banks track where the payments are

Siddharth Chandani (SC): We’re speaking to Santosh Tripathy, the practice lead for digital payments at SmartStream Technologies. Those artificial intelligence (AI) in payments, what perspectives on AI have changed since we started talking about some of its potential competitive advantages? How can we measure the impact? And how do we really separate the correct thing from the noise?

Santosh Tripathy (ST): We have been working in the AI area for quite some time. We were collaborating with our clients. We realised that you can develop anything independently, but if you work with the real use cases, with the clients, you'll have a great solution. What we did, we started talking to our clients globally and in different areas, like reconciliation, cash management, remittances, SWIFT’s global payments innovation (gpi) and we identified a couple of use cases. Those use cases were solved using our AI engine.

We are saying, “that's the way forward”. You can have a great AI engine, but it has to solve our use cases. In the last couple of years, that has happened. A couple of other players in this space have done similar work. They are collaborating and not working independently. That's the way forward. You have to collaborate, work together with the clients on the use cases and we’ll have great results. AI is incredible, but not on its own. It has to work with a solution.

SC: What in your opinion is making the adaption of AI harder for banks? Is it the people? Is it the legacy mindset? Or is it a combination of both?

ST: It's a combination of everything. Resistance to change is there. The changes are inevitable. We have to work with both realities. For sure, we’ll have resistance from people, but at the same time, the legacy software that large banks have is also creating a bottleneck. What we are seeing are new solutions that would have newer platforms. Anything that is new: cross-border remittances, digital payments or cash manager liquidity solutions, could be on your platform. That's a very positive sign we are seeing. They will have newer technologies.

Also, AI is very promising. Over a period of time, like it or not, people have to adapt AI. They have to start seeing this as an asset, not really as a threat. How about if you can achieve a 99.9% matching rate with AI and you can focus on something important? The risks are good for you, good for your bank, you can add value.

However, some of it has not been translated well or passed on to the users. Users have not been communicated with the benefits. Because of so much noise, it has been perceived as a threat to what they're doing, rather than a compliment. It's like, because we invented airplanes, trains ran out of business. What we have to do is we have to work with this user group and tell them the benefits. Once we have the solutions that solve the use cases and once they appreciate the benefit, we’ll not have this problem. It will happen in due course.

SC: There’s an emergence of these domestic real-time settlement infrastructures: New Payments Platform (NPP) in Australia, Fast and Secure Transfers (FAST) in Singapore and PromptPay in Thailand. The impetus on cross-border payments is much more. Where do you see the role of AI in terms of transformation?

ST: Real-time payment is also the future. At the moment, it's customer to customer. Eventually, it will become customer to business, business to customer and business to business. It will happen.

Large networks like Mastercard and Visa, they appreciated the challenge and they are collaborating with the technology companies. Mastercard acquired VocaLink, and now they are helping in this infrastructure of real-time payments. Real-time payments have been there in Singapore and marketing there for quite some time. The benefits have been realised. And we now know why we need real-time payments. Now you have to extend that from customer to the business, and then it will be great.

How do we use AI in cross-border remittances? Domestic is great, but cross-border remittances have been a big revenue stream for banks, and anyone can see the revenue aspect of it. Some people have also seen the challenge or the friction that is part of this whole cross-border remittances: multiple parties, multiple countries we are dealing with, different correspondents, regulations and domestic.

Certain countries are in the risk zone, some are blacklisted. In asset and liability management (ALM), the regulations that you have to deal with exist. What we could do is we can iron out some of those challenges with technology. AI is definitely a way forward, blockchain as well. Some companies like Ripple have taken over that idea and are moving ahead. Like Mastercard and Visa, we are seeing conversations around having a blockchain-based domestic scheme.

AI would stay. Real-time payments initiatives across the globe are there. In the Nordics, they are taking it beyond the country and doing it on a regional basis. In other countries, it's a country initiative. For Nordics, it's a regional initiative.

SC: Kindly share successfully implemented use case in areas of AI and reconciliation?

ST: United Kingdom (UK) has been a big market. We have many clients: large clients and small clients. What helps SmartStream is, because we have a large, diverse customer base, we are able to look at the use cases from a category one bank to a category two, or small emerging bank or digital-only bank. We have a client base across it.

In the UK, in terms of frequency relations, what we have done is we have implemented AI engine along with our reconciliation engine. Our reconciliation platform has been a global leader for a while. We are now going a step further, where we want to narrow down the unmatched and unreconciled risk item, even smaller percentage, and our customers have been very supportive. They have worked with us.

In our solutions, we already have AI implemented. In the UK, very soon, we'll have numbers to share with you. At the moment, you can see a very warm response we have gotten from our clients. One of the largest banks in the world is now using that engine with AI in a solution.

SC: It's been three years. Have we learned anything?

ST: A lot of thought has gone through it. SWIFT has been a very secure network for sure. The challenge was when the third-party vendor and the hackers are getting smarter. Their technology is great, but they have some easy ways. They just need to find that one entry point and then they enter into the SWIFT network. So the entry point was somewhere else but ultimately, remittances were impacted.

Now there’s stop and recall with real-time payments, with SWIFT’s gpi tracker – all of these are helping beneficial banks, settlement banks and correspondent banks to track where the payment is. If something goes wrong, stop and recall is great. The use cases are available, the rails are there, but even that is just one part of the solution.

The largest objective is, if you're working with banks, they have to interact with the platform or tracker, they have to update and they have to read. One option is to go and manually read from the tracker like a DHL Express. Second is if you keep getting notified on your cell phone or in your payment system over exactly where your payment is and if there is an alert, they’ll automatically notify you or send a message to the correspondent and talk.

We have a solution called advancement control, which is also working on SWIFT’s gpi. Our engine is interacting with their tracker, having all those workflows built in, so that the banks that are using the tracker get the added benefit of also automating the workflows. It's a great initiative, stop and recall. Additionally, because once you have real-time information, like if you withdraw money from your account, you immediately get a notification on your cell phone. If you did not use it, you'll get a notification and you’ll immediately call the bank and stop the payment, stop or block the credit card.

Likewise, what we'll see in remittances, they'll have real-time information of their payment flow. You'll also know who is supposed to pay for this remittance. This is going to increase the speed, the transparency and the security.

We’re very bullish about the initiative of SWIFT in this, but this is not the end. This is just the beginning. Once we get into that, we’ll realise the need of the customer is much more and SWIFT will realise it. It's a collaboration of all the banks so I'm sure that we are also part of that community. Everyone is talking. Everyone is giving their inputs. At SIBOS, payments have been a big topic. SWIFT’s gpi is leading the way.

SC: Thank you Santosh. 

ST: The pleasure is mine.