Transcript
Speaker: Welcome to HSBC Global Viewpoint, the podcast series that brings together business leaders and industry experts to explore the latest global insights, trends, and opportunities.
Speaker: Make sure you're subscribed to stay up to date with new episodes.
Speaker: Thanks for listening.
Speaker: And now onto today's show.
Speaker: Welcome to TMI's TreasuryCast.
Speaker: I'm Eleanor Hill, Editorial Consultant at TMI, and I'm delighted to be joined today by two special guests to discuss the future of AI with a special focus on payments.
Speaker: And those guests are Manish Pohli, who is Head of Global Payment Solutions at HSBC,
Speaker: and Amit Malik, who is Digital Payments Lead Europe at Accenture.
Speaker: So welcome both of you.
Speaker: It's so good to have you here, Manish.
Speaker: It's good to have you back.
Speaker: We've had you on a few times before.
Speaker: But you've just collaborated together to launch a very interesting report.
Speaker: It's called Navigating the AI Wave, Innovations in Commercial Payments.
Speaker: And of course, it's exploring the evolving role of AI, including Gen AI in Treasury and payments.
Speaker: But why would you say that this is the right time to focus on these topics?
Speaker: How do you see AI impacting the world of payments?
Speaker: Manish, maybe you can kick us off on that.
Speaker: Great.
Speaker: Thank you, Eleanor.
Speaker: First of all, it's a pleasure to be here on the podcast along with my good friend, Amit.
Speaker: I'm really looking forward to an engaging discussion.
Speaker: And talking about discussions, it's always interesting to speak about new technologies and their practical impact when discussing payments, and AI is no exception.
Speaker: And for us, it all starts with the clients.
Speaker: So out of our entire client base of about 1.3 million customers, there has been an increasing amount of engagement that we've had with clients who want to discuss AI adoption journey.
Speaker: So they don't want to just hear about the theory, but they want to hear about the practical.
Speaker: What is the industry doing?
Speaker: How is the technology manifesting itself?
Speaker: And when we talk to decision makers, so we did this HSBC Digital Horizon survey, 86% of decision makers have said that they expect generative AI to create opportunities for their business in treasuries and beyond.
Speaker: And the general theme, which is one of the reasons we put this report together to share our knowledge and to talk about what's happening in the industry, but the general theme is that AI is going beyond simple automation.
Speaker: and their opportunities for strategic applications in modern treasury and payment systems, a lot of which have not been fully harnessed to this stage.
Speaker: And the intent is there to have this conversation, this discussion, and share best practices.
Speaker: And we said in our report that over the next three years, finance leaders expect AI tools to become increasingly integral to the risk management and decision-making processes.
Speaker: with about 60% of them saying that AI expected to be either very or extremely useful as compared to 40% today.
Speaker: So we felt that this is the right time to have this conversation and share some of our thinking in this space.
Speaker: Brilliant stuff, Manish, and some really good stats there.
Speaker: I think it does bring it to life.
Speaker: And every treasurer I speak to is asking me about AI, honestly.
Speaker: So I think you're bang on the money with that.
Speaker: But Amit, what would you add to that?
Speaker: First of all, thank you for inviting me here and really excited to be a part of this interesting conversation.
Speaker: What we are seeing is a few different things in the market.
Speaker: First of all, I think AI today is really accessible.
Speaker: It's no longer a complex technological construct, a black box.
Speaker: People understand AI better and it's more easily accessible.
Speaker: Two is we are also seeing guardrails getting set up around AI.
Speaker: So there are do's and don'ts around AI.
Speaker: There are ethical considerations around AI.
Speaker: There are regulatory considerations around AI.
Speaker: So that's important.
Speaker: Competition is increasing within the space.
Speaker: Disruptors are building newer innovative propositions, more efficient propositions using AI.
Speaker: So organizations which don't embrace AI
Speaker: are likely to get left out in this entire process.
Speaker: And all of this is evident from a latest Accenture survey, which we did, where 74% of the respondents said that they have seen tangible benefits from AI within the last 12 months in terms of their investments.
Speaker: And 60% additionally said that they would continue significant investments into AI through 2026 and on.
Speaker: So now is the right time.
Speaker: Wonderful stuff.
Speaker: Manish, do you agree now's the right time?
Speaker: Absolutely it is.
Speaker: If I can just add, from our perspective, AI is not new to us at HBC and we've been exploring this technology since more than a decade ago.
Speaker: And, you know, we established our AI center of excellence in 2010.
Speaker: But like Amit said, like, you know, we're now seeing more practical use cases emerging and we envisage the technology to be applied more into payment solutions.
Speaker: and helping create more value for our clients.
Speaker: So this ties in with the theme of accelerating innovation, which is very important for us and is actually core and central to the dialogue that we're having with clients.
Speaker: So we're committed to supporting our clients through this transformative journey, bringing best ideas that we have along with what we are hearing in the industry.
Speaker: And that's why this collaboration with Accenture is actually important for us in terms of sharing our knowledge here.
Speaker: Yeah, and like you say, it's been going on for a while.
Speaker: You've not just started looking at AI recently.
Speaker: The bank has a long history of innovation, likewise Accenture.
Speaker: So it's a good partnership.
Speaker: But when we think about AI and particularly the way that it's disrupting payments, quite often we like to think about consumer payments, retail payments, but our audience would love to know what's going on in the commercial payments industry.
Speaker: How is that being disrupted by digital innovators?
Speaker: Maybe you can give us some examples, especially like how AI is actually reshaping payments and treasury.
Speaker: And is the disruption all about AI?
Speaker: Should we be thinking about other things as well?
Speaker: Again, Manish, maybe let's come to you first.
Speaker: Great, Eleanor.
Speaker: And that's a very interesting question.
Speaker: So maybe let me zoom out a little in terms of disruptions in commercial payments, and I'll go deeper into AI after that.
Speaker: And what we have been seeing that over the last few years, there has been an acceleration in payments and in treasury functions.
Speaker: If I had to kind of put a finger on the top two or three ones that I can think about, one is clearly real-time payments, a topic that's been explored quite extensively.
Speaker: And one that we believe that it is not about just sending payments 24 by seven with enriched data, but it is about helping our customers build new businesses and build new experiences.
Speaker: So I'd say, you know, real time payments has continued to be a super important avenue for innovation in commercial payments.
Speaker: The second is a wider adoption of APIs across industries.
Speaker: And this is again, not just about making your own processes digital, but it's about connecting various participants in an ecosystem.
Speaker: And finally, the whole form factor of money in terms of stable coins and digital assets, which has proven that you can move money much faster, 24 by 7, and even cross-border at the speed of light.
Speaker: The world of commercial payments has continued to innovate, and we expect this innovation journey to kind of further accelerate.
Speaker: And if you see common themes over here, they'd be in the theme of retail experiences influencing commercial payments.
Speaker: and also commercial payments being focused around improving each individual aspect of an attribute of a payment, which is either cost, which is speed, which is transparency, or security or convenience.
Speaker: It is about ensuring that the innovation that is happening is absolutely real and impactful to common practices and problems that our customers face.
Speaker: Now, zooming into AI, which is where your question more was, AI is no exception.
Speaker: And we're seeing the adoption of AI in commercial applications very similar to retail applications.
Speaker: I'll give maybe two or three simple examples to make it real.
Speaker: First is, we've seen a lot of payment fintechs that are using AI-powered products
Speaker: receipts at the checkout experience.
Speaker: So when shoppers make a purchase, they receive a receipt, which not only tracks the transaction, but also uses predictive knowledge to inform where future preferences, purchase preferences from that merchant could be.
Speaker: And that is something which is at the intersection of both the consumer experience as well as the business experience.
Speaker: Our own partner, Feedzine, which uses AI and machine learning to enhance fraud prevention and detection.
Speaker: That's technology we are embracing ourselves because specifically with the rise in the velocity of money, it is important for us to use new technologies to help us prevent fraud by identifying suspicious activities and also reducing false positives because you want to balance this with experience.
Speaker: And finally, in our own estate, what we have been using is launching capabilities like FX Prompt and said simplistically, where if a customer is making a cross-border international payment, we use the intelligence that we have built by processing billions of transactions in determining the currency of a particular account and using that information to give insights to customers about what is the most optimal way for them to do FX and
Speaker: and manage the whole cross-currency payment.
Speaker: And this is actually powered entirely by our AI engine.
Speaker: So my summary point would be, Eleanor, there's rapid acceleration of using technologies for commercial payments.
Speaker: And now we think that more of that is going to happen with AI being at the center of those disruptions.
Speaker: Absolutely.
Speaker: And they're so concrete and so real, those use cases.
Speaker: I know the FX one will have prompted a lot of listeners to have a little think about that and come and chat to you.
Speaker: That's something that people have been asking for for a long time, those kind of capabilities.
Speaker: So exciting stuff.
Speaker: And I look forward to hearing a little bit more about that as it goes on.
Speaker: But let's bring you back in.
Speaker: What are your thoughts on this?
Speaker: First of all, Manish, thank you.
Speaker: Excellent use cases and very good examples of how we are seeing usage and adoption of AI within payments and especially commercial payments.
Speaker: Perhaps to reinforce some of your thoughts, we are seeing companies like Stripe significantly invest in fraud-based AI tools, using and harnessing data from multiple customers around the world to provide instant fraud messages to merchants who use Stripe solutions embedded at a point of sale.
Speaker: Very relevant, very useful.
Speaker: Similarly, we are seeing significant impact of AI within the customer servicing space.
Speaker: MasterCard has launched an AI service agent, which is supposed to significantly increase the efficiency of customer services provided by MasterCard to both businesses and consumers all over the world.
Speaker: Similarly, Klarna has launched an AI bot, which today replaces the job of 700 customer service agents.
Speaker: So we are talking about very, very scaled examples of efficiency within the world of AI in payments.
Speaker: At Accenture, we are piloting a range of AI-based innovations to support payments customers, especially in technology and operations.
Speaker: The first example I can think about is creating AI-based components to support fraud detection, especially authorized push payment fraud.
Speaker: Additionally, within payments, we have this challenge of content moderation and preventing abusive messages in payment fields, which will become more and more common with ISO, expanding the number of fields available within payments.
Speaker: And AI is doing a great job of producing that and understanding the intent behind the payment and the messages sent as a part of the payments field.
Speaker: We are also looking at automatic generation of content, especially as organizations move from old empty standards to the modern ISO 2022 standards within the payments field.
Speaker: So all in all, a lot of interesting and very, very practical and relevant use cases coming up within the world of AI, especially in payments.
Speaker: Yeah, Amit, wow, thank you so much.
Speaker: And I think a lot of the listeners won't have even considered anything like abusive messages, like you say, in these longer payment fields.
Speaker: And there's just so much to consider and so much potential for AI.
Speaker: But Manish, let's come back to the reports.
Speaker: I mentioned the name of it at the beginning.
Speaker: So it's Navigating the AI Wave, Innovations in Commercial Payments.
Speaker: You've
Speaker: collaborated on it together.
Speaker: I've had a very good read, had a sneak peek, thankfully, for this podcast.
Speaker: You've got loads in there in the report about the value opportunity for Gen AI, particularly looking at payments and treasury management.
Speaker: So let's speak a little bit about that.
Speaker: Yeah, sure, Eleanor.
Speaker: You know, in our report, we tried to make this real from a payments and treasury perspective of where we're seeing near-field value opportunity for generative AI.
Speaker: One is on client servicing.
Speaker: We're increasingly seeing generative AI becoming more and more advanced in terms of interpreting basic client inquiries and presenting data in a structured and accessible format.
Speaker: And when you pair that with automation workflows, which most banks and we have specifically been investing in, we can combine these two technologies to considerably shorten turnaround time and deliver better client experience.
Speaker: And what I really like about this is that
Speaker: On top of that, it actually takes away the dependency of the human in doing the first round of data analysis.
Speaker: And it just brings in a level of consistency.
Speaker: It brings in a level of language independence.
Speaker: It just helps in us not just being prompter, but just being responsive to customers in a more superior way.
Speaker: The second example that I probably would point out is around payment healing.
Speaker: And all of us in the world of payments know and are familiar with transaction failures and discrepancies and delays that happen due to various factors.
Speaker: These could be technological challenges.
Speaker: This could be compliance issues.
Speaker: This could be human errors along the way.
Speaker: And while machine learning historically has been used to automate the identification of the potential friction points,
Speaker: Well, we are seeing an increased use cases on generative AI that can help create synthetic data, fill in gaps in transaction records, but also go back to the root cause of the error so that they can be avoided in the future by either improved learning within our own operation shops or us upstreaming the learning into client organizations so that the data comes in in a format that is cleaner and supportive of more straight-through processing.
Speaker: Amit, let's come over to you and maybe you could give us your thoughts as well on what we're seeing in terms of Treasury and AI and how it can shift the way that we've done things traditionally and all of those pain points that Treasurers have spoken about for so long.
Speaker: The traditional use case within Treasury has been cash flow forecasting.
Speaker: But recently, over the last 12 to 18 months, we are seeing an evolution and maturity of that use case evolving from cash flow forecasting to actually end-to-end liquidity management.
Speaker: And what I mean by that is Treasurers today will be able to use AI to do liquidity management across the entire supply chain, which means not only within their organization, but with external partners and other players within the ecosystem.
Speaker: And the way this is manifesting itself in what we are calling as Treasury 2.0 is these things.
Speaker: One is we are able to get today much better treasury insights around the performance of an organization with additional drill downs into individual performance of business units and using Gen.AI to build a narrative and reporting around that.
Speaker: So that's number one.
Speaker: Number two is within the cash flow forecasting space, we are seeing
Speaker: much more efficient cash flow forecasting we are seeing significant reduction in the amount of money spent in supply chain finance and we are also seeing increased availability of working capital and better cash concentration available and thirdly i think all of this is evolving into what we are thinking as a booking capital control tower where people will have a control tower in place where
Speaker: working capital view across the entire organization is very, very visible.
Speaker: And this results in significant improvement in various processes, including collections and the amount of time spent supporting customer service centers or contact centers.
Speaker: So quite a lot going on within this space.
Speaker: And I think within the world of commercial payments, AI and Gen AI will have a significant impact in how we do treasury today and in the foreseeable future.
Speaker: Yeah, and I love the idea of the working capital control tower.
Speaker: And I'm sure a lot of the listeners will love that as well and be thinking about how they can go about implementing that for themselves and working towards that.
Speaker: But Manish, when we talk about AI, of course, it's
Speaker: super tempting to focus on how transformative it is, but we also have to think that it's not necessarily a silver bullet.
Speaker: So tell me a little bit about what you think its limitations are and also what practical steps can treasurers take to maximize the value of AI within those limitations.
Speaker: Yeah, sure, El-Nor, and you're absolutely right.
Speaker: This is not a silver bullet.
Speaker: It is a very transformative technology.
Speaker: There are a few things that have to be right for AI deployment to be effective.
Speaker: The first is, there is a considerable amount of dependency on high quality data.
Speaker: AI systems only thrive on well-structured data.
Speaker: And with the data is poor and inconsistent, the outputs will be mildly inaccurate to be widely inaccurate.
Speaker: And we've said the same, we've talked about this in our report also, where our survey had shown that 82% of respondents agree that supporting data collection and analysis is a top priority and even a precondition to adopting AI.
Speaker: So data dependency is one big theme that I would talk about.
Speaker: And sticking to the point of data, I'd also say data residency restrictions, where data is stored, where it can be processed,
Speaker: can complicate the ability of AI to be used effectively.
Speaker: And again, as a global organization, these are challenges that we deal with.
Speaker: And I know our clients also deal with similar challenges because of the wide distribution of data and data legislation, specifically evolving legislation.
Speaker: And these restrictions can hinder organizations from accessing the necessary data to be able to train the AI models and ultimately impact the performance and reliability of those applications.
Speaker: And the final thing is a word of caution that I would express is that as we are building and as our customers build AI models and AI solutions, we need to be aware and beware of the over-reliance on AI.
Speaker: So there needs to be a specific regulated industry like ours.
Speaker: There needs to be a good balance between AI and human oversight.
Speaker: And while AI can solve many of the operational challenges that are there, it cannot replace human expertise or judgment in areas that require creativity or nuanced decision making, which again, in the world of financial services and banking, we are dealing with every single day.
Speaker: Yeah, absolutely.
Speaker: And I know the treasurers will be thinking the same.
Speaker: They might want suggestions from AI around where they should invest their surplus cash, but ultimately they still want to make the decision themselves.
Speaker: So, yeah, so important, that human element.
Speaker: Amit, what would you add on to that?
Speaker: How would you expand?
Speaker: I think the first is explainability and transparency.
Speaker: Most AI-based systems today, especially those using deep learning,
Speaker: are almost like black boxes and sometimes it is difficult to explain why the AI took a particular decision or made a particular recommendation.
Speaker: Example, in Treasury, a transaction can be earmarked as fraud without clearly identifying why has it done so.
Speaker: So I think use of AI along with explainability of the AI is very important and organizations need to take measures to do that.
Speaker: So that's number one.
Speaker: Talking about the data dependency, AI systems are fundamentally restricted by the fact that they can be biased based on the data input they have received.
Speaker: And this can create multiple ethical and ethical concerns.
Speaker: For example, if you're looking at a lending AI,
Speaker: it might not make the same credit offer to a particular demographic based on biased data.
Speaker: So there is a need to start thinking about the data dependency within AI.
Speaker: These are additional important considerations if we are to make a success out of Gen AI and AI.
Speaker: Yeah, absolutely.
Speaker: And that bias, like you say, there's so much to consider, even as treasurers looking at
Speaker: They're looking at credit on suppliers, buyers, country risk, et cetera.
Speaker: There's a lot to consider if they've got AI models looking at those sorts of things.
Speaker: So great that we flagged that.
Speaker: But Manish, I wanted to come back to the practical steps that treasurers can take.
Speaker: What would you say for them today?
Speaker: Because it can be quite daunting, I think.
Speaker: So where should they start?
Speaker: You know, the practical advice we give is, one is focus on modernizing your platforms and making them more digital.
Speaker: And that is AI with no AI and no regret bet that
Speaker: all treasurers should be focused on.
Speaker: And that allows for, in the context of AI, allows for seamless data integration and also creates good strong foundations for once you get outputs out of an AI engine and a solution, the deployment of those becomes more and more effective.
Speaker: The second I would say is it is very important to have a very robust governance framework for two reasons.
Speaker: One is
Speaker: It helps channel the energies of an organization to use cases that are most useful and where AI can be the most effective technology.
Speaker: And I think that is sometimes just selecting those use cases can become a challenge.
Speaker: And I think having a good, efficient governance framework is helpful for that.
Speaker: And secondly, is it helps in terms of making sure
Speaker: that many of the challenges that Amit just talked about around ethical concerns and bias, there's good oversight on them.
Speaker: And finally, I would always say that it is important for all organizations to start building frameworks or collaborating externally.
Speaker: There is so much change, rapid change in pace and technology that's happening, that if you're not exposed yourself to external collaboration,
Speaker: there's always a risk that what you build today is outdated by the time you complete the build.
Speaker: And I think it's a very, very important consideration.
Speaker: In some cases, it doesn't come naturally to many organizations that are probably focused on just building internally without an external view.
Speaker: Yeah, absolutely.
Speaker: And like you say, you then get the cutting edge of what's going on if you're working with people that have the budget and the expertise to be at the cutting edge.
Speaker: And it's very useful to have a sounding board sometimes as well and sort of check where you're going with everything.
Speaker: Amit, what would you add from your side in terms of practical steps that treasurers can take here?
Speaker: Yes, I can think of three practical and simple steps which treasurers and payment leaders within various organizations can start thinking about.
Speaker: One is value targeting, which means that they need to understand what do they expect out of AI?
Speaker: What are the targets they are setting for themselves and start quantifying it so that they can measure it?
Speaker: The next is value realization, which means that they need to set up a value realization office or in simpler words, a program office to start delivering those programs as per those targets and timelines they have set for themselves.
Speaker: The third thing to think about is value sustainability.
Speaker: which means that this is not a one-off exercise.
Speaker: This has to be done periodically and repeatedly to get the ongoing and continuous benefits, and they need to be cognizant of that.
Speaker: So I think if they can focus on those three simple steps, they will start deriving the benefits we all talked about during this conversation.
Speaker: Wonderful stuff.
Speaker: Thank you so much for that.
Speaker: We have had a brilliant conversation so far.
Speaker: We've given the listeners so much to think about, I know.
Speaker: But I would love to kind of wrap up with just some concluding thoughts from each of you.
Speaker: Key takeaways for the audience.
Speaker: Maybe I can come to you first.
Speaker: We have talked about so many exciting use cases.
Speaker: We are almost talking of a world of payments in which there is
Speaker: no fraud or zero fraud, in which transactions get settled in real time, money moves in real time.
Speaker: And we're also talking about a world where businesses, big and small, will have access to a range of financial management and treasury management tools.
Speaker: All of this will only happen if we take responsibility, if we use AI with wisdom and integrity.
Speaker: My concluding thoughts are that we should think about not what AI can do for payments, what AI should do for payments.
Speaker: Brilliant.
Speaker: I love it.
Speaker: Thank you.
Speaker: And like you say, yes, there's a big responsibility that comes with this technology as well.
Speaker: Manish, what would you add to that?
Speaker: We all need to recognize that we're in a very early stage.
Speaker: of generative AI usage.
Speaker: And as we talk about in our report, we have a lot of optimism about what the future holds.
Speaker: And it is always important to remember that AI is a transformative tool, but it is not intended to replace humans.
Speaker: So we have the opportunity over here to integrate a human's creativity, ethical judgment, and contextual understanding with AI's capability to process massive data.
Speaker: And when you put both of them together, that empowers us to make informed decisions, enhance customer experience,
Speaker: and navigate complex challenges in the payment space that we have today and we will continue to have.
Speaker: So I think it's just staying tuned and close to this space.
Speaker: And from our perspective at HSBC, we'll continue to publish thought leadership papers as the industry and the use of technology in the space evolves.
Speaker: I think it's very important that we all stay connected and very close to this technology because it is genuinely transformative.
Speaker: Absolutely.
Speaker: And I love what you said about it's human plus AI, not human versus AI.
Speaker: And that's so important to think about as we go forward.
Speaker: And like you mentioned, the thought leadership there.
Speaker: So just to everyone listening or watching, please do go and have a look for this latest white paper.
Speaker: If you head to the HSBC Treasury Solutions Group website, there's actually a whole AI series there.
Speaker: And you can also download the report there as well.
Speaker: But for
Speaker: now thank you both ever so much for joining us today it's been really good fun chatting to you hearing all about working capital control towers payments healing you've introduced us to some amazing new terms and opened up a lot of possibilities in treasurer's minds i think so thank you very much thank you for joining us at hsbc global viewpoint we hope you enjoyed the discussion make sure you're subscribed to stay up to date with new episodes
