Become a Creator today!Start creating today - Share your story with the world!
Start for free
00:00:00
00:00:01
Treasury Beyond Borders - Treasury’s AI future: Balancing efficiency, trust and human judgement
821 Plays25 days ago
Artificial intelligence is starting to make a tangible impact in treasury, from sharper forecasting to faster decision-making.
In this episode of Treasury Beyond Borders, Rahul Badhwar, Global Head of Corporate Sales, Markets and Securities Services, HSBC and Allen Li, Global Head of Digital and Distribution, Markets and Securities Services, HSBC discuss how treasurers can benefit from AI today, and where its limits remain. They debate the role of AI in automating workflows, improving data analysis, and supporting risk management – while underlining the importance of governance, transparency, and human judgement. The episode also introduces HSBC AI Markets, a platform that brings data, analytics, and execution together for more efficient treasury operations.
Disclaimers:
HSBC AI Markets is for existing clients. The information contained in marketing materials is for demonstration only. The Bank is not responsible for any loss, damage, or other consequences of any kind that you may incur or suffer as a result of, arising from or relating to your use or reference of the information in this material. The information provided to you on HSBC AI Markets is intended only for professional sophisticated and/or eligible clients, investors or counterparties.
HSBC AI Markets allows users to access certain HSBC data using Natural Language Processing (‘NLP’) which provides the ability to interpret and comprehend human language. The NLP parser employed by HSBC AI Markets will attempt to match a user’s query to the appropriate answer based on the data available to it. Aspects of this NLP model use Machine Learning which is a subset of artificial intelligence using mathematical tools and algorithms to create a model that can be used to make predictions. HSBC AI Markets does not use Generative AI which builds on Machine Learning to take content in the form of text or other formats and generate new context as text or other formats. Information provided by HSBC AI Markets is indicative, its accuracy may vary and should be used for information purposes only.
Recommended
Transcript
Introduction to 'Treasury Beyond Borders'
00:00:01
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.
00:00:13
Speaker
Make sure you're subscribed to stay up to date with new episodes. Thanks for listening, and now onto today's show.
00:00:22
Speaker
Welcome to the special TMI podcast series in collaboration with HSBC called Treasury Beyond Borders, Balancing Risk and Opportunity Through Global Growth. I'm Eleanor Hill, Editorial Consultant at TMI.
00:00:35
Speaker
And over the course of this series of podcasts, I'll be exploring how global treasury is evolving amid geopolitical shifts and digital transformation. And across the episodes, with insights from HSBC experts, we'll dive into key themes both globally and across different regions, from managing risk in uncertain markets to leveraging new trade corridors and emerging technologies.
The Role of AI in Treasury Management
00:00:57
Speaker
And on that note, today we're here to talk about the very hot topic of ai We'll talk a little bit generally about what's going on, and then we're going to dive into HSBC's AI markets. So very exciting, lots to talk about. And we have a returner to the podcast. First of all, of our two guests, we have Rahul Badoir, who is Global Head of Corporate so Sales, Markets and Security Services, HSBC.
00:01:20
Speaker
And he is joined by Alan Lee, who is Global Head of Digital and Distribution, Markets and Security Services, also HSBC. So welcome guys, it's great to have you here. Rahul, I'm so pleased you decided to come back and it wasn't too awful last time.
00:01:36
Speaker
No, and and not not at all, Ellen. And before we kick off, I must say I've yet to find an AI model that can duplicate your energy and enthusiasm. Even though you're talking about AI, for me, the human context is still very, very important.
00:01:48
Speaker
Absolutely is. Let's dive into it, guys, because we've got so much to talk about. So let's have a little bit of a view on the current landscape. And maybe you could give us a sense on the pain points that AI actually solving for treasurers today, because I know there's a A lot of hype out there, so it'd be good to get a sense of maybe the reality of what it's actually doing. So Rahul, I'm going to pick on you first for a bit of a macro view, if that's okay.
00:02:13
Speaker
Absolutely. So if you think about AI, I think what it has done brilliantly is helped with automation of tasks. And linked to that automation is enhancement of data analytics, because obviously AI has the ability to go through volumes and volumes of data at a much faster pace than a human being would.
00:02:32
Speaker
And therefore, from a treasurer's perspective, I think what it really does is it arms the treasurers with much better quality input when it comes to risk management, when it comes to hedging.
00:02:43
Speaker
So, for example, if you're a treasurer and you are taking decisions on cash flow hedging, you know hedging your future payables or your future receivables, forecasting that future payable or future receivable with some degree of accuracy is very important. yeah Otherwise, you risk being under hedged or over hedged in relation to your hedge ratio.
00:03:03
Speaker
Now, I think where AI has really played an important role is improving the quality and the accuracy of that forecasting. So that has definitely been a very good tool in the arsenal for the corporate treasurers.
00:03:15
Speaker
I think also with AI analysis of historical data, when you're looking at either foreign exchange markets or interest rate markets or just macro variables, I think that is also very important because obviously as a treasurer, it's very important to get a sense of how the underlying currency and interest rates and macroeconomic variables could be behaving in the future.
00:03:36
Speaker
And again, you know, using that in the hedging decisions by analyzing this data, again, using AI, I think becomes very, very relevant.
AI's Analytical Strengths and Limitations
00:03:43
Speaker
And I think last but not least, if you want to quantify at risk by, again, analyzing vast volumes of data to understand what are correlations, right?
00:03:53
Speaker
What are dependencies that can exist between different asset classes, What are some of the patterns that do emerge? I think these kinds of things, AI is very, very good at. Where I think there is still room to improve, I guess is you know, just the gut feel that a human being brings to the table, which is obviously currently infer not replicated by ei And to me, a lot of that gut feel is having lived through an experience, whether it was a global financial crisis, whether it was how the markets behaved when the dot-com bubble burst or more recently, you know, what happened with the pandemic.
00:04:28
Speaker
I think that gut feel is very important. But the one cautionary word, you know, when it comes to us in the financial services space is that we are a part of a regulated industry. And the therefore, for me, the success of AI depends on making sure that any AI outputs that we deliver,
00:04:46
Speaker
treat our customers fairly, add value, and actually respect both the letter as well as the spirit of the law. So a whole bunch of positive, but also a little cautionary tale in terms of how we need to approach things.
00:04:59
Speaker
I think that's such a a good summary there, Rahul, of everything that's going on. There's a lot of promise for treasurers, especially in that hedging space. I know ah few who are looking at the FX exposure forecasting piece in particular. I'm very excited about that, but I'm absolutely with you on the gut feel. And people are, some people are bit funny about gut feel, but it's so much more than instinct. It's all of these little risk signals that your brain has picked up and catalogued over time through going through experiences. So yeah, lots to do there and completely agree on the governance angle as well. And the ethics, it's not just about the letter of the law, it's about the spirit of it as well.
AI and Process Automation in Treasury
00:05:36
Speaker
But Alan, let's bring you in because I know you've got loads to share as well. Maybe you can tell us a bit more from the tech side of things, what's going on.
00:05:44
Speaker
Very quickly, I think building on Rahul's point, I think at the moment, the power of AI in Treasury is about process automation and it's about predictive analytics also.
00:05:56
Speaker
I think that then helps to deliver smarter decision and give meaningful cost reduction via greater efficiency and accuracy. Yeah. specifically, know, from a technology point, it's important to realize that at this time, AI is not just a buzzword.
00:06:13
Speaker
You know, unlike probably many that we have been through. Many that I've written about, many that we've probably spoken about over years, but I completely agree with you. And it's not as new as people think either, even though there's a lot of recent breakthroughs going on.
00:06:28
Speaker
And I think it's currently driving real change across almost every single aspect workflow, including treasury. And I think ah some of those recent breakthroughs in generative AI are are expectably transformative. And I want to talk about maybe four yeah know from a technical point of view. I think the first one is about how we can help with information retrieval and discovery.
00:06:54
Speaker
And then there's another one which is about, you know, how it helps people to synthesis information together and as a result, facilitate the research and the learning. It's also important to recognize how in many of those use cases out there now, it helps people to create digital asset.
00:07:13
Speaker
And in that, I mean, whether it's content or in some cases about software development, you build a dashboard, you build your application. It can be done within minutes in instead of people spending hours or days and weeks to create something equivalent previously.
00:07:29
Speaker
And some of the top you know revenue generating applications has been in this content creation and application software
Challenges in AI Integration
00:07:37
Speaker
development space. yeah I think last but not least, we start to see a lot is about around human and computer interaction because essentially generative AI you know allow you to take natural language processing to the next level.
00:07:52
Speaker
And that was a lot of questions, but also thinking ah about what exactly the next generation of interaction between human and all those are service digital capabilities is going to be.
00:08:06
Speaker
And these to me are changes that are fundamental. Now, that being said, I do believe that a fully autonomous big treasury system is still aspirational at this stage. And a lot of research we have seen out there, you know probably predicting about, you know, between 15 to 30% of like autonomous decision-making in probably a few years' time. So I think at this stage, we're still not quite there to get a to fully autonomous system.
00:08:37
Speaker
And you know some of the key gaps, I think Rahul had already mentioned just now, and they are about you know security because with great power also comes a risk that you need to manage.
00:08:49
Speaker
It's also about data privacy and governance. The regulatory compliance space is evolving really, really quickly. And you know last but not least, big change also comes with the need of change management.
00:09:02
Speaker
And in that, that's also the talent gap that we need to make sure that our and employee, our people are you know given the right training to get it to the right level of skills to adapt and to make the best use of this new transformative technology.
00:09:18
Speaker
So in short, I would say that AI is revolutionizing how trearys operate today, but to realize the full vision will require continuous investment and governance and skill set.
00:09:33
Speaker
Absolutely. And that people just so, so important. and It's not about just focusing on the technology. It's about bringing people along at the same time. I think it's really important for our audience to understand a bit more about how they need to change their treasury operating models fully benefit from AI because it's not just plug and play. There's various things that you need to rethink to actually make the most out of it.
00:09:57
Speaker
You've got to look at systems. You've got to look at people. Like you said, there's also a mindset piece. So Rahul, maybe you can give us your thoughts on that and what people actually need to think about changing so that making the most of the AI opportunities that are there.
00:10:10
Speaker
Yeah, Elena, you know, so the question is always, is it systems that need to change? it people or is it mindset? And, you know, sometimes good things always come in threes. So for me, it there' is actually a combination of all these three things, you know.
00:10:24
Speaker
So first and foremost, I think companies need to be putting an emphasis which drives a mindset change with regard to AI adoption. I think in many ways, we are the cusp of a cultural change. and change is never easy.
00:10:38
Speaker
I think we all know that. But, you know, my view was we need to go into this ah with a glass, half-full kind of view. And when I look around, you know, I see that AI is making people far more productive than they could otherwise be.
00:10:52
Speaker
And therefore, we can do more with the help of AI and then truly focus our time on more value-added tasks. And for me, that's also in a way, you know, improves quality of life because you're not spending hours and hours in the office doing, you know, routine stuff, which can be automated and and delivered with extremely high quality, if not much better quality than a human could.
00:11:13
Speaker
So AI is helping us reinvent ourselves. It's something that we need to embrace. But apart from the human view and the mindset, I think this is all about systems and tools and how AI moves away from just analyzing historical data to actually learning and and and making predictions about the future.
00:11:33
Speaker
I think that is going to be really, really critical as things evolve. Because right now, you know analyzing the past is great, but how do we analyze the past and start getting AI to give us guidance on yeah What are some of the future scenarios that could in unfold?
00:11:49
Speaker
And why is it that one scenario is more likely than the other? That I think is really going to be a game changer in new Course. or an awful lot to look forward to and sort of experiment with. But like you say, it's about making sure that we're making the most of it ourselves.
00:12:05
Speaker
It's very easy to get left behind. but I think there's an awful lot of opportunities for treasurers here. And we've spoken before on our previous podcast about all of the uncertainty that's out there at the moment. It's a very turbulent landscape, especially treasury and trade.
00:12:20
Speaker
So how do you see AI helping treasurers there? How is it going to be a game changer for them in this kind of environment that we're dealing with at the moment? So, you Elena, with everything that is happening around the world, I think the job of a treasurer has become far more complex.
00:12:37
Speaker
I think gone are the days where, you you know, majority of the treasurers essentially operated in their home markets and therefore essentially had to deal with very domestic interest rate and currency issues. so Today, as corporates are diversifying supply chains, are expanding into new markets in search of you know new customers for their product or service, what goes hand in hand with this expansion is that you're dealing in a far greater number of currencies.
00:13:04
Speaker
You're having to think about funding your subsidiaries, which means you're probably dealing in far more interest rates.
AI in Currency and Risk Management
00:13:10
Speaker
And all this requires expertise. yeah And what is very important is then how AI can help our corporate treasurers to actually manage and deal with increasing number of currencies and interest rates in their portfolio.
00:13:26
Speaker
Because again, this is a lot about data. yeah and and And consuming their data and and understanding the microeconomic variables, understanding the regulatory framework, right? In emerging markets, the regulatory framework is very different from in some of the developed markets.
00:13:40
Speaker
And I think increasingly, companies are also looking at ways at becoming more efficient. You know, they want to centralize risk management in many cases. And, you know, put all the power behind ah centralized team rather than having 10 different teams in 10 different markets around the world.
00:13:55
Speaker
I think increasingly the solution is not necessarily throwing more people at the problem, but using AI to be far more efficient. And this empowering a centralized team to take effective decisions because you've got ai doing a lot of the heavy lifting for you, yeah which is tactical. But the strategic decision making falls on the corporate treasurer and their team.
00:14:17
Speaker
Yeah. And it's that refrain, you know, helping the treasurer do more with less, which we've been talking about for so many years, but now AI is actually here and able to help with that at long last. So yeah, an awful lot to look forward to around that efficiency and gaining time back for those strategic tasks. And I know the bank has been busy looking at AI and how it can help. So Alan, let's talk a little bit about HSBC AI markets. We've mentioned this in a couple of the podcasts in this series already, but for anyone who hasn't listened to those yet, A, go and have a listen to them.
00:14:50
Speaker
But B, Alan, maybe tell us a little bit about HSBC AI markets. What kind of scenarios would it help treasurers in, in terms of improving decision speed or accuracy?
00:15:01
Speaker
Give us a sense of how it works. Sure. I ah will start with the overview about the value of position here for Treasury. So um if you look at HBC AI markets, right, it try to elevate that decision-making by combining the proprietary data we have in HBC, couple that with but advanced analytics, but not only that, you also have real-time execution liquidity on the platform, and you put all these together into one single platform in one single portal.
00:15:33
Speaker
And that's where we leverage the power of technology to make the workflow easier for our user. And to give a few examples relevant to the day in the life of people in treasury, so take cash yield and money market optimization.
HSBC's AI Markets Platform
00:15:50
Speaker
Now, what you have when you bring all these capabilities within one single place is that it's not always about trading risk for them. You know, sometimes also about, you know when they have cash left in the account, they want to find the best way, the most efficient way to get the investment return on that cash.
00:16:08
Speaker
Now, having all those high-quality data, but also the pricing information with real-time liquidity in one place means that they can very, very easily scan through all the different opportunities and then find the one that's optimized for their niche and preference.
00:16:26
Speaker
So that is about capturing the full potential of that cash balance. The next item I would say is slightly more technical, but it's about the ability for our user to customize a dashboard that is tailored for their needs.
00:16:42
Speaker
If you look at, you know someone coming in whose focus might be on a certain currency pair risk or, you know whether it's on cash management or looking at the global exposure.
00:16:53
Speaker
They can, on the platform, quite easily configure a single dashboard and containing critical analytics, events, calendar, trading tool, or even scenario models.
00:17:03
Speaker
And you can then share that board with your colleague in your team, ah adjusting daily for different priority, and then collaborate in real time. So I think this personalization support both individual decision-making, but also team-based workflow is quite important for large multinational or fast-moving SME treasury team.
00:17:25
Speaker
Yeah. Oh, I know our listeners will be very interested in that customizable dashboard sound of things. And like you say, being able to share it with colleagues and lots of insights at their fingertips. But Alan, one of the things treasurers worry about a little bit with AI is how much they can trust the outputs. I know much of that is really around the the gen AI side of things. We're not talking about gen AI. We're talking about Natural language processing, AI that's been around for longer and has been used more and tested more.
00:17:58
Speaker
But just give us a sense, a bit of reassurance around how much we can trust the outputs from this tool, how explainable they are, how audit proof they are, especially under the regulatory scrutiny that's out there at the moment.
00:18:12
Speaker
Yes, sure. And this is probably going to get slightly more technical, but as Rahul had mentioned, in a heavily regulated industry and the trust and the transparency are fundamental for us.
00:18:24
Speaker
And this has been achieved within HBC AI markets through a combination of technology choice and robust governance. And there are few things that I would like to share.
00:18:35
Speaker
The first one is, to your point, the output is deterministic and non-generative AI. Now, for all critical functions you can see in HBC AI markets, our AI tick deliver deterministic and non-generative AI results.
00:18:52
Speaker
So this means that all the output you get would be consistent, would be traceable, and it would be based strictly on those approved database ah we have from HBC and our models.
00:19:05
Speaker
There won't be any unpredictable and unexplainable variants. so yeah The second thing a around the clarity is that when people get the result back, we want to make sure that the result get back is self-contained and it's a clear response.
00:19:22
Speaker
When we want to make sure that the result we get back, we clearly indicate what that result is for so that that a result, that response itself is self-contained and it's clear in itself.
00:19:33
Speaker
I think the other thing is about all the good practice about the governance and the documentation. So in all the way from the data, the design, the release, the development, each and every single process, we need to make sure that it's well documented because if not, then it should be treated as an unknown database.
00:19:55
Speaker
So following those rigorous governance framework and that there are lots now in the market, you know, especially given the development of generative AI and the importance of data. We follow those process when it comes to data, documentation, design, and development you know with the highest standard to make sure that this is done properly.
00:20:15
Speaker
ah Last but not least, the model we are using is proprietary NLP model that has been developed over the past eight years. So this allows us the capabilities, but also with our own design process to ensure that the result, the design is done in a way that is fit for purpose. Because when you have a generic model, it might not necessarily understand a lot of the domain experts are driving use in the financial market. So this proprietary model also is quite important for us for our continuous and iterative improvement to enhance that user experience, but also the accuracy we need for our use cases. so
00:20:56
Speaker
The last point I want to make is that the reason why those data are ah special, are proprietary, and they are high quality is that a lot of those analytics is backed up by real liquidity.
00:21:08
Speaker
and Take, for example, you if you want to get analytics on central bank pricing, which gives you a sense about a potential central bank rate change now.
00:21:19
Speaker
Those data is highly valuable because those data behind the scenes is deriving from the liquidity we're providing to the market. And that account by liquidity is critical items to ensure that quality of the data is out there.
00:21:34
Speaker
you yeah because you can get all the technical expect right. But if you don't really have that level of data with that level of depth in the quality to back it out, it might not be a data and of the quality people need to understand about the market. So,
00:21:50
Speaker
In summary, I would say it's not just a digital interface and it's one that's fully governed by a fully governed suite of HBC digital capabilities, designed from ground out to be trusted, explainable, and audit-ready for any level of scrutiny.
00:22:10
Speaker
Oh, awesome stuff, Alan. Thank you for that. I can feel your passion for it coming through. and I love that you called it special. ah it feels like it has its own personality. You spent an awful lot of time and developing the technology and like you say, making sure it's just right for everyone and has all of the right parameters and depth and quality of data.
00:22:29
Speaker
But Rahul, I wanted to bring you back in because we've had Alan in his comfort zone there talking all about the tech, which I know the listeners will have enjoyed. But I'd love to leave everyone just with One key takeaway, something to think about. So for our audience, our global treasurers out there who maybe dabbled a tiny bit in AI, but they want to do something a bit more meaningful.
00:22:52
Speaker
What's the first step, would you say, that's actually going to move the needle rather than just frittering around the edges?
Strategic Use of AI in Business
00:22:59
Speaker
Eleanor, I'm probably going to lean on some quotes which some very prominent leaders of tech and companies have given in the past in relation to AI.
00:23:07
Speaker
but And one of them talked about that the future of AI is about augmenting human capabilities. And the other talked about, or rather said, that AI will not replace humans, but those who use AI will replace those who don't.
00:23:22
Speaker
Yeah. So my one message is this is coming back to the same thing. It has to be torn from the top. We've got to realize that humanity is at the cusp of a revolution.
00:23:32
Speaker
Just as how the birth of the Internet changed humanity in many ways, the birth of AI is going to change the way we do our jobs and how we define productivity. So one is torn from the top and embracing it.
00:23:45
Speaker
Secondly, just start with identifying tasks that are repetitive, that rely on analyzing large volumes of data, and start thinking about what is it that you can do? Because as companies are growing and as the world is becoming more complex,
00:24:00
Speaker
The only way to deal with this efficiently is to use as AI to take over some of the tasks that were being done in the past by human beings, and then really use the human talent and human capital for the strategic brain power that a human being brings to the table.
00:24:16
Speaker
And I think that's a great marriage and bonding between AI and using human talent. It really is. And there's so much opportunity out there. And it's a great chance for people to improve themselves, I think, as well. You know, the people that take an interest in it. There's so much you can learn.
00:24:32
Speaker
i've done Absolutely. I thought my learning curve was coming to a flat thing more than then I talked to Alan and I'm like, OK, I'm back in school again. So, so yeah, it it's steepened my learning curve all over again.
00:24:44
Speaker
Thank you so much for coming on the podcast today. It's been a real pleasure and chatting about AI, which is one of my favorite topics and hearing the passion from your side. so thank you ever so much. That brings us to the end of this episode of Treasury Beyond Borders. Thank you everyone for tuning in. I hope you enjoyed the discussion as much as we did.
00:25:03
Speaker
Please be sure to subscribe and stay updated on our other episodes and watch out for the articles accompanying this series as well. But for now, thanks again to our wonderful speakers and thanks to everyone for listening.
00:25:17
Speaker
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.