WEBVTT 1 00:00:02.096 --> 00:00:12.982 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. 2 00:00:13.803 --> 00:00:16.364 Make sure you're subscribed to stay up to date with new episodes. 3 00:00:16.705 --> 00:00:17.385 Thanks for listening. 4 00:00:17.525 --> 00:00:18.886 And now onto today's show. 5 00:00:36.447 --> 00:00:39.229 Hello and welcome to TMI's Treasurycast. 6 00:00:39.410 --> 00:00:42.152 I'm Eleanor Hill, Editorial Consultant at TMI. 7 00:00:42.613 --> 00:00:50.219 And today's episode forms part of our special Redefining Treasury series produced in collaboration with HSBC. 8 00:00:50.600 --> 00:00:56.085 We're exploring how the Treasury function is evolving in response to new technologies, new expectations and... 9 00:00:56.625 --> 00:01:00.367 and the very rapidly changing operating environment out there. 10 00:01:00.507 --> 00:01:04.189 So for today's episode, we're looking at the very hot topic of connectivity. 11 00:01:04.249 --> 00:01:05.590 So there's loads to cover. 12 00:01:05.670 --> 00:01:16.416 We'll be looking at how treasury teams can make sure they've got the right information and connectivity capabilities and everything in the right place at the right time, which is perhaps easier said than done. 13 00:01:16.796 --> 00:01:20.618 But we have two experts here to help us along through this topic. 14 00:01:20.638 --> 00:01:33.002 So I'm delighted to be joined by Amber Henderson-Smott, who is Global Head of Client Connectivity, and Peter Crawley, Global Head of Corporate Sales, both within Global Payment Solutions at HSBC. 15 00:01:33.042 --> 00:01:35.123 So Amber and Peter, welcome. 16 00:01:35.283 --> 00:01:36.604 It's great to have you here. 17 00:01:36.624 --> 00:01:37.684 I hope you're doing well. 18 00:01:37.844 --> 00:01:44.107 I know we've got absolutely loads to get into today, and it's such a good topic because it's moving so, so quickly. 19 00:01:44.167 --> 00:01:44.347 But... 20 00:01:44.927 --> 00:01:56.330 Amber, let's kick off maybe with a look at best in class connectivity, what that looks like today, how it's changed for Treasurers, because I think there's so many different expectations now. 21 00:01:56.390 --> 00:02:07.932 It's not just the faster movement of data and payments, but as I said in the intro, it's getting that right information, the capabilities at the point where they need them, when they need them. 22 00:02:08.032 --> 00:02:10.573 So, Amber, what are you seeing in terms of that best in class? 23 00:02:11.013 --> 00:02:11.613 Thanks, Eleanor. 24 00:02:11.794 --> 00:02:12.614 Great to be here. 25 00:02:12.754 --> 00:02:16.437 Look, treasurers definitely have more complexity than ever before to manage. 26 00:02:17.297 --> 00:02:25.283 As you said, payments are moving faster, market dynamics are quicker, but also new forms of digital money are growing and they're definitely here to stay. 27 00:02:26.003 --> 00:02:32.750 I think at the same time, we're also hearing from our clients that they want leaner operations, they want tighter controls and better technology. 28 00:02:33.070 --> 00:02:37.835 And at the end of the day, all of this is driving greater demand for greater precision. 29 00:02:37.856 --> 00:02:44.983 So what this really means is getting the right information at the right time to our clients so they can act on it with confidence. 30 00:02:45.283 --> 00:02:48.327 And that's what best in class connectivity is really all about. 31 00:02:48.867 --> 00:02:57.459 You know, bank connectivity needs to be able to enable faster decision, you know, process, you know, automation and adapt for the future. 32 00:02:57.599 --> 00:03:04.589 So whether that's ISO, AI, digital currencies or whatever next innovation is coming in financial infrastructure. 33 00:03:05.830 --> 00:03:09.972 And many of our clients are also on their own technology journeys to streamline their treasury. 34 00:03:10.572 --> 00:03:19.916 Last year, we ran a treasury pulse survey, and that told us that just over 40% of our clients are using an ERP or TMS today to support their processes. 35 00:03:20.576 --> 00:03:24.218 But it also told us about 30% are still relying on spreadsheets. 36 00:03:24.318 --> 00:03:30.881 So, you know, there's definitely no single best connectivity model that's right for everyone or every use case. 37 00:03:31.521 --> 00:03:37.723 And so at HSBC, our focus is really about bringing multiple connectivity options together in a single experience. 38 00:03:38.184 --> 00:03:46.407 So treasurers might use Swift for high volume payment processing, whilst also using APIs to access real-time balances or instant credit notifications. 39 00:03:47.207 --> 00:03:56.657 I think there's even more value unlocked from bank connectivity in a treasury context when it's tightly integrated into the systems and workflows our clients already use. 40 00:03:57.077 --> 00:04:00.000 So it could be an ERP or it could just be Microsoft Excel. 41 00:04:00.701 --> 00:04:07.908 And that gives treasurers better access to information, greater automation and ultimately a much faster path to make decisions and take action. 42 00:04:08.408 --> 00:04:18.391 Yeah, but like you say, Amber, within the systems that they're already using and not it doesn't have to be some huge transformation project, which I'm sure we'll get into, but very interesting stats there from that survey. 43 00:04:18.571 --> 00:04:27.473 And I'm sure a number of the listeners will feel the pain on the spreadsheet side of the 30 odd percent, as you said, Peter, anything you'd add to that? 44 00:04:28.012 --> 00:04:28.912 Yeah, morning, Eleanor. 45 00:04:28.932 --> 00:04:29.612 Thanks. 46 00:04:29.792 --> 00:04:35.174 And I think Amber gave a great summary there, but perhaps I could just give one example to bring it to life. 47 00:04:35.254 --> 00:04:39.615 And we have a customer bandwidth, cloud communications platform as a service company. 48 00:04:39.635 --> 00:04:44.576 And what's interesting about their story is it wasn't about just implementing another bank connection. 49 00:04:44.656 --> 00:04:52.158 It started with a conversation about what the treasury team was trying to achieve from a visibility control of the cash. 50 00:04:52.978 --> 00:04:58.860 ability to make decisions based on real-time information rather than waiting for files and reports. 51 00:04:58.880 --> 00:05:02.282 So we kind of worked alongside them in their chosen treasury platform. 52 00:05:02.302 --> 00:05:03.582 In that case, it was Travata. 53 00:05:04.122 --> 00:05:09.424 And we used APIs to connect kind of real-time balance and transaction data directly into the systems they use. 54 00:05:09.484 --> 00:05:18.748 And so I think that meant kind of moving away from a model where information arrived, sort of scheduled batches, more into when treasury needs it. 55 00:05:18.808 --> 00:05:19.068 And so 56 00:05:19.808 --> 00:05:24.091 I think a lot of this is also about, you know, going beyond simply seeing balances foster. 57 00:05:24.851 --> 00:05:37.300 You know, when treasury teams have immediate access to transaction level information, they can reconcile more efficiently and identify exceptions, improve cash flow and, you know, make funding and payment decisions with confidence as well. 58 00:05:37.320 --> 00:05:44.865 So, you know, I think this is really important to set the foundation here around, you know, scalable connectivity solutions for the future. 59 00:05:45.627 --> 00:05:46.327 Absolutely. 60 00:05:46.627 --> 00:05:50.089 And like you said, when Treasury needs that data, it's there. 61 00:05:50.149 --> 00:05:53.770 So it might be actually an on demand model rather than a real time model. 62 00:05:53.810 --> 00:05:57.751 And there's all these different ways that are going to work for different clients. 63 00:05:57.811 --> 00:06:00.232 But top marks, Peter, for bringing us an example. 64 00:06:00.292 --> 00:06:02.333 That's what we always ask our podcast guests. 65 00:06:02.353 --> 00:06:04.213 They're like, please, can you bring it to life for us? 66 00:06:04.233 --> 00:06:08.315 So well done for squeezing that in so early on in the podcast. 67 00:06:09.795 --> 00:06:15.019 Peter, I'm going to pick your brain a little bit more in terms of what corporates are asking for these days. 68 00:06:15.059 --> 00:06:18.142 I know, as we've said, there's no kind of one size fits all. 69 00:06:18.382 --> 00:06:19.263 Everyone is different. 70 00:06:19.683 --> 00:06:25.608 But when you're having these conversations around connectivity with clients, who's shaping those requirements now? 71 00:06:25.788 --> 00:06:27.250 Is it Treasury alone? 72 00:06:27.430 --> 00:06:28.771 Is there a lot of input from the CFO? 73 00:06:29.312 --> 00:06:30.614 Are the tech teams involved? 74 00:06:31.054 --> 00:06:32.897 Is it even the wider business? 75 00:06:33.157 --> 00:06:42.711 And how are you seeing priorities differ according to the organization type, the operating model, and also that level of digital maturity? 76 00:06:42.731 --> 00:06:44.694 Just give us a sense of what you're seeing out there. 77 00:06:45.174 --> 00:06:46.535 Yeah, I mean, that's a great question. 78 00:06:46.575 --> 00:06:47.376 And there's a lot in that. 79 00:06:47.436 --> 00:06:52.639 I mean, what we're seeing is that the agenda is no longer set by Treasury alone. 80 00:06:52.659 --> 00:06:59.844 I mean, Treasury remains at the center because they're responsible for liquidity and risk and funding and, of course, banking relationships. 81 00:06:59.924 --> 00:07:06.128 But increasingly, they're working alongside, as you say, CFOs, but also technology teams, other business leaders. 82 00:07:06.649 --> 00:07:11.452 And these stakeholders within the corporate organizations all have a stake in how cash and money. 83 00:07:11.912 --> 00:07:14.596 financial data flows through the organization. 84 00:07:15.517 --> 00:07:22.806 What's common across almost every client conversation is the desire for better visibility and foster decision-making. 85 00:07:22.967 --> 00:07:27.032 You know, the environments our clients operate in are often volatile, 86 00:07:27.532 --> 00:07:29.493 The markets can move quickly. 87 00:07:29.813 --> 00:07:30.533 They're growing. 88 00:07:31.074 --> 00:07:36.616 Organizations want real-time view of liquidity rather than relying on yesterday's information. 89 00:07:36.656 --> 00:07:47.261 And they want to understand where cash sits globally, forecast more accurately and make funding and investment and hedging decisions, as I said earlier, with a bit more confidence. 90 00:07:47.461 --> 00:07:48.881 But it's not just that visibility. 91 00:07:48.921 --> 00:07:52.023 They're looking to also simplify complexity. 92 00:07:52.323 --> 00:07:55.104 I think you touched on models. 93 00:07:55.164 --> 00:07:57.005 Organizations expand internationally. 94 00:07:57.545 --> 00:08:10.259 They can often inherit multiple banking relationships, sometimes different connectivity models and have quite fragmented processes that have almost been imported as a function of their growth. 95 00:08:10.939 --> 00:08:16.886 And there's this growing focus we see on standardization in addition to this automation. 96 00:08:17.466 --> 00:08:20.129 And often that comes with establishing some form of innovation. 97 00:08:20.349 --> 00:08:21.931 single connectivity framework. 98 00:08:21.951 --> 00:08:33.043 I use the word framework because it's sometimes, you know, a basket of different tools underneath that connectivity framework that they can then scale across their business lines and different markets. 99 00:08:33.912 --> 00:08:38.453 I mean, the priorities do depend on the digital maturity of the customer as well. 100 00:08:38.873 --> 00:08:46.995 I mean, more advanced organizations are kind of focused on APIs and real-time data and how AI can help automation in decision making. 101 00:08:47.515 --> 00:09:03.558 But there are many, many others that are still earlier in their transformation journeys, concentrating on, say, standardizing processes or improving their controls or reducing manual intervention, which still exists in the payment workflows and cash management workflows. 102 00:09:04.298 --> 00:09:08.801 I think regardless of where they are, the direction of travel is remarkably consistent, though. 103 00:09:09.381 --> 00:09:11.303 And they want connectivity simpler. 104 00:09:11.323 --> 00:09:16.286 You know, they want it to be more intelligent, more integrated into the systems that they use every day. 105 00:09:16.886 --> 00:09:18.367 And then I just want to access data. 106 00:09:19.548 --> 00:09:23.610 And I haven't yet mentioned it, but they also want this to be done in a cost efficient way. 107 00:09:25.311 --> 00:09:26.372 Have your cake and eat it. 108 00:09:26.392 --> 00:09:27.273 That's absolutely it. 109 00:09:27.413 --> 00:09:27.553 Yeah. 110 00:09:27.573 --> 00:09:27.773 Yeah. 111 00:09:28.594 --> 00:09:29.494 Good stuff. 112 00:09:29.514 --> 00:09:32.017 Just to maybe add to what Peter said, right? 113 00:09:32.077 --> 00:09:37.321 I think, you know, we talk a lot about business and operational teams wanting simplification. 114 00:09:37.361 --> 00:09:41.665 They want things streamlined, automated, more embedded, less manual work. 115 00:09:41.745 --> 00:09:50.193 And so, you know, it's not surprising that technology is therefore becoming a really important part of Treasury or it's already there, but it's becoming much bigger. 116 00:09:50.813 --> 00:09:54.817 And so what that means is our clients' technology teams are now 117 00:09:55.217 --> 00:09:59.579 have a seat at the table when we are presenting on RFPs or having conversations. 118 00:09:59.619 --> 00:10:02.921 And maybe that wasn't the case three or four years ago. 119 00:10:03.041 --> 00:10:12.345 And so it's really important for us to make sure when we're looking at connectivity with customers, what is the right architecture, the security, the scalability? 120 00:10:12.385 --> 00:10:15.187 These are all things that are more important than ever before. 121 00:10:15.827 --> 00:10:19.130 And we also need to consider, as Peter said, you know, the cost, right? 122 00:10:19.170 --> 00:10:25.556 What's the internal effort needed to implement and then maintain each connection with the bank as it goes forward? 123 00:10:25.656 --> 00:10:33.963 And so for us at HSBC, it's not just a case of the connectivity itself, but also what's the developer and the implementation experience, right? 124 00:10:34.003 --> 00:10:35.744 These are just as important. 125 00:10:35.804 --> 00:10:43.651 And so, you know, good connectivity or best in class connectivity as we started off the session really needs to be easy to test, easy to deploy, right? 126 00:10:44.011 --> 00:10:45.312 and adapt and scale. 127 00:10:45.812 --> 00:10:58.060 And so I guess in practice that means things like quality API documentation, testing sandbox, and proven integrations with ERPs and TMS platforms such as SAP, FIS, Kariba. 128 00:10:58.501 --> 00:11:01.823 And all of that helps reduce the effort involved in connecting to the bank. 129 00:11:02.824 --> 00:11:07.847 To bring it to life, a good example is the recent work we did with Shantar Fork Jewellery Group in Hong Kong. 130 00:11:08.527 --> 00:11:15.132 We actually worked with them as they implemented their SAP S for Hana and also multi-bank connectivity with HSBC. 131 00:11:15.652 --> 00:11:22.797 And that helped them create a standard proven integration that they're actually now able to repeat across the entities as the business grows. 132 00:11:23.278 --> 00:11:28.021 And it's also given the group much better visibility and control over their payments and cash positions. 133 00:11:28.701 --> 00:11:29.902 Ah, brilliant stuff, Amber. 134 00:11:29.922 --> 00:11:30.662 Thank you for that. 135 00:11:30.782 --> 00:11:34.604 And again, gold star for the example for us for bringing it to life. 136 00:11:34.664 --> 00:11:35.365 You're so right. 137 00:11:35.465 --> 00:11:42.728 It's got to be, you know, it can't just work for the bank or be the best thing that the bank has seen in a different organisation. 138 00:11:42.748 --> 00:11:44.069 It has to work for the client. 139 00:11:44.149 --> 00:11:47.651 It has to be scalable and achievable for the client and maintainable. 140 00:11:48.071 --> 00:11:55.755 And all of those things pile into an equation that's going to be slightly different for everyone, but you can bring those experiences together. 141 00:11:56.462 --> 00:11:58.203 Brilliant stuff, Amber, thank you, Peter. 142 00:11:58.243 --> 00:12:03.045 I'm gonna come back to you because we've mentioned all of these connectivity ways. 143 00:12:03.065 --> 00:12:06.367 We've had to host the host, we've mentioned Swift, we talked about APIs. 144 00:12:06.807 --> 00:12:14.050 Of course, you mentioned AI and we've got these emerging agent to agent models that are coming as well, very exciting. 145 00:12:14.130 --> 00:12:21.714 But if you're a treasurer out there listening, how do you then choose the right approach for a particular process or outcome? 146 00:12:21.754 --> 00:12:24.175 Because it's like being in a sweet shop at the moment. 147 00:12:25.153 --> 00:12:27.854 Yeah, that's a good analogy, Eleanor. 148 00:12:28.194 --> 00:12:37.376 I think the starting point is recognizing that there isn't a single connectivity model that's right for every organization or every process. 149 00:12:37.397 --> 00:12:41.518 I mean, the question shouldn't be, which channel should I use? 150 00:12:41.558 --> 00:12:45.659 It should be rather, what outcome am I trying to achieve? 151 00:12:46.259 --> 00:12:58.767 I mean, if a treasurer needs real-time visibility of cash positions, balances or transaction activity embedded directly into their ERP or treasury management systems, APIs are often the natural choice. 152 00:12:59.448 --> 00:13:04.852 And they provide this kind of immediacy that we've referenced and access to information to support. 153 00:13:05.612 --> 00:13:23.531 decision-making, but then for higher volume payment processing, often the traditional channels like host-to-host connections and, and Swift remain incredibly important because they're proven and, and scalable and deeply embedded in many ways with many organizations operating models. 154 00:13:23.551 --> 00:13:23.812 Yeah. 155 00:13:24.332 --> 00:13:30.415 I mean, banking portals and mobile channels, you know, also continue to play a critical role. 156 00:13:30.656 --> 00:13:36.079 I mean, no matter how automated an organization becomes, there are always situations. 157 00:13:36.199 --> 00:13:42.662 I often get calls with something on a Friday that's urgent and has dropped out of the process. 158 00:13:42.742 --> 00:13:48.746 And, you know, we need to help investigate an exception or get something, you know, cross border really quickly. 159 00:13:49.246 --> 00:13:52.328 And so those channels are also needed as well. 160 00:13:52.368 --> 00:13:52.528 But... 161 00:13:53.148 --> 00:13:58.170 I'd say what we're increasingly seeing is that leading treasury organizations don't choose one model. 162 00:13:58.390 --> 00:14:00.570 You know, they combine several. 163 00:14:01.271 --> 00:14:09.773 They use connectivity channel where it delivers the most value creating almost an architecture that balances real time insight and ops resilience. 164 00:14:10.674 --> 00:14:16.756 Because that resilience is really important in large organizations, especially, and the scalability and cost. 165 00:14:17.536 --> 00:14:19.718 And the other factor is digital maturity. 166 00:14:19.738 --> 00:14:33.009 I mean, some organizations are already exploring AI-enabled workflows and agent-to-agent interactions, while others are focused on automating core processes and reducing manual intervention. 167 00:14:33.089 --> 00:14:35.251 I mean, both can be the right strategy. 168 00:14:35.271 --> 00:14:43.577 I think it comes down to what depends on where the business is on its transformation journey, which I've mentioned earlier. 169 00:14:44.058 --> 00:14:46.280 And really, we see our role at HSBC is... 170 00:14:47.055 --> 00:14:57.183 helping our clients navigate those choices and designing a sort of connectivity strategy, as it were, that supports today's priorities, but also tomorrow's ambitions. 171 00:14:57.563 --> 00:15:06.430 And that's why we continue to invest in the connectivity ecosystem, including newer digital capabilities that I'm sure Amber's going to get into. 172 00:15:06.490 --> 00:15:07.731 I don't want to front-run those. 173 00:15:07.791 --> 00:15:15.317 But ultimately, to your question, I mean, the best connectivity model is the one that fits the client's operating model today while giving them some form of... 174 00:15:15.977 --> 00:15:22.240 flexibility to evolve into the newer technologies and capabilities as their businesses grow and mature. 175 00:15:22.660 --> 00:15:24.020 Yeah, that makes sense, Peter. 176 00:15:24.060 --> 00:15:24.481 Thank you. 177 00:15:24.541 --> 00:15:40.427 Amber, how would you build on that in terms of choosing the right connectivity model and those implementation considerations that determine whether that connectivity is actually going to work successfully in practice, which I know is not always, doesn't always go as smoothly as we want. 178 00:15:40.868 --> 00:15:41.308 Yeah, so 179 00:15:42.048 --> 00:15:53.539 I think, as Peter said, right, there's a huge amount of complexity or it can feel really complex and almost overwhelming if you look at the different options that you have as an organisation and a lot of different considerations. 180 00:15:53.579 --> 00:16:01.427 I think the good news I like to share is there is a pretty straightforward implementation approach that we use, which is really four steps. 181 00:16:01.927 --> 00:16:04.769 The first is the outcome, which is really important. 182 00:16:04.809 --> 00:16:08.912 The reason we're all here is to deliver for our clients on the outcomes they're looking for. 183 00:16:09.312 --> 00:16:11.734 The second step is matching the right connectivity. 184 00:16:11.874 --> 00:16:12.915 The third is testing. 185 00:16:12.995 --> 00:16:14.756 Then the fourth is scaling safely. 186 00:16:15.437 --> 00:16:20.480 And so if we think about the outcome, you know, it's usually going to be streamlining a particular treasury task. 187 00:16:20.921 --> 00:16:23.963 So it could be payment initiation and approvals. 188 00:16:24.503 --> 00:16:31.826 It could be delivering real-time balances into the TMS, or it might just be starting with automating statement reconciliation. 189 00:16:32.387 --> 00:16:43.152 And so the first step really is to identify where is information fragmented or where manual intervention exists, or maybe where addressing any delayed action could give the greatest benefit. 190 00:16:43.952 --> 00:16:50.255 Once you're clear on the outcome that you want to go after, the next step is really matching the right bank connectivity to the outcome. 191 00:16:50.855 --> 00:16:57.618 And so again, you know, without repeating the points that Peter said, this is where treasurers need to think about what's the speed of information they need? 192 00:16:58.078 --> 00:17:00.599 What kind of transaction volumes are they looking at? 193 00:17:01.179 --> 00:17:03.780 Is the process batch based or event driven? 194 00:17:04.320 --> 00:17:07.581 And then they also need to think about where does the work need to happen? 195 00:17:08.181 --> 00:17:17.645 Obviously, ideally within the ERP would be preferable, but obviously, depending on the level of automation and control they're looking to achieve, that'll be a key factor. 196 00:17:18.305 --> 00:17:22.608 And so the message really is you don't need to transfer everything at once. 197 00:17:23.129 --> 00:17:25.750 We recommend building on one use case first. 198 00:17:25.871 --> 00:17:29.673 So again, it could be a single process, a payment type, a market. 199 00:17:30.294 --> 00:17:34.517 And this is really important because the implementation is not just about technical readiness. 200 00:17:35.257 --> 00:17:39.960 There's often changes that our clients will need to make internally to their operational processes. 201 00:17:40.460 --> 00:17:45.183 There could even be changes to certain roles and responsibilities within the Treasury team. 202 00:17:45.203 --> 00:17:55.229 So starting small, then scaling allows you to work out any wrinkles so you can then speed up the implementation with more confidence across any remaining processes, entities and markets. 203 00:17:56.276 --> 00:17:58.037 Some brilliant suggestions there. 204 00:17:58.137 --> 00:17:59.598 Amber, thank you very much for that. 205 00:17:59.678 --> 00:18:02.439 I wanted to just diversify a little bit. 206 00:18:02.459 --> 00:18:07.502 We've touched on a lot of the interesting shifts that are happening in the market out there. 207 00:18:08.122 --> 00:18:09.783 We've mentioned AI, etc. 208 00:18:10.123 --> 00:18:18.388 But also seeing more banking information and services appearing directly within ERP and TMS workflows. 209 00:18:18.788 --> 00:18:20.470 So tell us a little bit about that. 210 00:18:20.570 --> 00:18:25.556 How far has embedded banking in inverted commas progressed? 211 00:18:25.756 --> 00:18:27.698 Where do you see it creating value? 212 00:18:27.939 --> 00:18:35.747 And then how are banks giving clients greater choice, but without requiring a new integration every time that needs change? 213 00:18:36.348 --> 00:18:36.848 Yeah, great. 214 00:18:36.868 --> 00:18:46.152 Look, I think I said earlier that just over 40% of our clients are now using an ERP or TMS with some level of host-to-host or API connectivity behind the scenes. 215 00:18:46.712 --> 00:18:51.294 And so embedded banking's made real progress, but there's definitely more work to do. 216 00:18:51.994 --> 00:19:03.038 The value of embedded banking really comes from being able to bring balances, transaction details, payment initiation, and exception management directly into the client's own platform. 217 00:19:03.799 --> 00:19:14.490 And so that reduces the need to move between systems or bring information together manually, so improves the visibility while strengthening controls, which is definitely what our treasurers are looking for. 218 00:19:15.677 --> 00:19:22.161 It also allows bank data to be brought together and sit alongside or be aggregated with internal data. 219 00:19:22.401 --> 00:19:34.067 So this could be internal forecasts, payables, policies, maybe internal approvals, which gives the Treasury team much more context that helps them make faster, better informed decisions. 220 00:19:34.787 --> 00:19:39.810 And that same context will be really important as AI becomes more widely used in Treasury. 221 00:19:39.890 --> 00:19:41.391 And I think we're going to chat about that. 222 00:19:41.751 --> 00:19:42.932 a little bit more coming up. 223 00:19:43.552 --> 00:19:47.134 So the benefit's not just about optimizing the day-to-day processes. 224 00:19:47.314 --> 00:19:50.716 The other benefit is about shortening the onboarding time. 225 00:19:51.156 --> 00:19:58.179 So reducing bespoke integration work and having to train treasury teams multiple times on different bank portals. 226 00:19:58.700 --> 00:20:04.883 And so again, you know, building on, you know, what I said before, it makes the implementation a lot more repeatable and predictable. 227 00:20:05.663 --> 00:20:10.268 I spoke earlier about Chowtay Fook's use of SAP multi-bank connectivity. 228 00:20:10.768 --> 00:20:17.195 Another really good example of embedded banking is our pre-integration with FIS that we launched at the start of the year. 229 00:20:17.615 --> 00:20:25.142 So this gives clients a proven integration of real-time balances and transaction information through our FIS integrity, quantum and payment hub. 230 00:20:25.843 --> 00:20:32.645 We also launched for some of our smaller customers the HSBC Integrator that has established connections to NetSuite and Sage. 231 00:20:33.285 --> 00:20:42.608 And these proven integrations and established connections have helped our existing HSBC customers ultimately enable their connectivity in a matter of days instead of what used to take weeks. 232 00:20:44.246 --> 00:21:03.448 busy times on your side by the sounds of it good stuff um and we look forward to learning more about all of these integrations as we go but peter i wanted to come to you because i'd love to learn a little bit more about the real-time connectivity piece because when i speak to some treasurers who maybe aren't ready for it yet 233 00:21:03.868 --> 00:21:09.794 They don't necessarily see the benefits of what is in their mind just getting the same information sooner. 234 00:21:10.295 --> 00:21:14.279 But tell us a bit about how it can actually change your business outcome. 235 00:21:14.339 --> 00:21:17.141 I know you've got a little case study up your sleeve. 236 00:21:17.181 --> 00:21:27.392 I think it's a Unilever one around kind of immediate credit notifications that impacts receivables, available credit, and then the company's ability to sell a little bit more. 237 00:21:27.951 --> 00:21:42.323 Yeah, I mean, firstly, what I'd say, and you're absolutely right, when I speak with treasurers, immediacy is often synonymous with getting payments out earlier, which might work against the working capital agenda that they have. 238 00:21:43.044 --> 00:21:48.028 But I think this is where the conversation moves beyond connectivity and into business value. 239 00:21:48.869 --> 00:21:50.771 And real-time connectivity isn't simply... 240 00:21:51.311 --> 00:21:55.313 about receiving the same information, as you say, a few hours earlier or instantly. 241 00:21:55.393 --> 00:22:00.375 It's helping change the decisions a company can make and the outcomes it can achieve. 242 00:22:00.755 --> 00:22:04.777 And, you know, the Unilever case is a fantastic example of that. 243 00:22:04.797 --> 00:22:12.901 I mean, historically, their cash flow forecasting relied on end of day reporting across multiple time zones, which meant treasury teams were often 244 00:22:13.741 --> 00:22:17.783 making decisions using information that was already out of date or quite old. 245 00:22:18.423 --> 00:22:24.526 Funding and liquidity and risk decisions were often being made on a retrospective view of the business. 246 00:22:25.527 --> 00:22:38.313 And working with HSBC, Unilever implemented API connectivity that provides real-time visibility across more than, around about 300 accounts spanning Asia and Europe. 247 00:22:39.213 --> 00:22:43.656 Treasury can now see balances and cash movements as they happen, regardless of the market and time zone. 248 00:22:44.357 --> 00:22:48.299 But the real crux of it is what's important, is what it enables, really. 249 00:22:48.800 --> 00:22:50.561 Incoming funds are identified immediately. 250 00:22:51.101 --> 00:22:55.865 The Treasury gets confidence and gains confidence in having that liquidity available sooner. 251 00:22:56.645 --> 00:23:05.999 That means with the reconciliation of that, which is happening much faster, that credit limits can be released more quickly to distributors and trading partners. 252 00:23:06.019 --> 00:23:10.325 And that allows orders to be fulfilled and goods to be shipped sooner. 253 00:23:10.685 --> 00:23:12.828 So in other words, like better connectivity doesn't just... 254 00:23:13.469 --> 00:23:18.693 improve treasury operations, it can actually help accelerate revenue generation and support business growth. 255 00:23:19.294 --> 00:23:38.229 You know, the same visibility also improves the forecasting, funding decisions, working capital management, and treasury teams can deploy the cash as a result more efficiently and avoid having, you know, the unnecessary buffers that are synonymous with, you know, batch store and released style cash management arrangements. 256 00:23:38.810 --> 00:23:41.392 So, I mean, the real breakthrough, you know, 257 00:23:41.532 --> 00:23:44.533 is that connectivity is no longer just an information channel. 258 00:23:44.553 --> 00:23:48.355 It becomes an enabler of better and faster decisions. 259 00:23:48.976 --> 00:24:04.763 And when a treasurer can act on information, the moment it becomes available, the benefits in my mind extend far beyond the treasury function into the wider organization's ability to serve customers and grow revenues and, of course, manage their risks as well. 260 00:24:05.185 --> 00:24:06.285 Yeah, absolutely. 261 00:24:06.385 --> 00:24:09.687 It's definitely a ripple effect in a very large way. 262 00:24:09.727 --> 00:24:15.329 And I know the Treasury team at Unilever very well, and they're always looking for market leading solutions. 263 00:24:15.349 --> 00:24:17.790 They're always very quick to adopt things. 264 00:24:17.870 --> 00:24:21.351 So definitely one for our listeners to have a look out for. 265 00:24:22.171 --> 00:24:25.913 Do follow what the Unilever team are up to because they've always got good ideas. 266 00:24:27.553 --> 00:24:37.042 I know I've spoken to Unilever a little bit about AI as well and pluses and minuses and you've got to be cautious, but you've also got to take the opportunity. 267 00:24:37.143 --> 00:24:39.024 So let's delve into AI a bit more. 268 00:24:39.084 --> 00:24:45.831 You mentioned the kind of agent to agent workflows that we are starting to see coming through or people are at least exploring them. 269 00:24:46.331 --> 00:24:58.040 What are you seeing clients prioritizing as they introduce AI into their treasury workflows, whether it's they're looking at forecasting, whether they're thinking about fraud detection, liquidity, execution? 270 00:24:58.140 --> 00:24:59.341 Tell us what you're seeing. 271 00:25:00.042 --> 00:25:01.563 Yeah, all of the above. 272 00:25:01.703 --> 00:25:08.808 I mean, it's such an exciting space, a really big topic as well, and very much came alive this year, I would say. 273 00:25:08.828 --> 00:25:14.633 At the beginning of the year, it seemed to have sort of popped right onto the top of the agenda of CFOs and treasurers. 274 00:25:15.373 --> 00:25:19.914 Treasury teams are prioritizing AI, where I would say it delivers measurable outcomes. 275 00:25:19.954 --> 00:25:24.636 And you've referenced a few examples, all of them we've seen coming to life. 276 00:25:24.976 --> 00:25:26.496 I mean, it's already entering Treasury. 277 00:25:26.996 --> 00:25:32.838 A quarter of the Treasurers that we survey say that AI is either in production or in a pilot. 278 00:25:33.898 --> 00:25:36.019 and another third roughly are evaluating it. 279 00:25:36.059 --> 00:25:39.800 So I think there is an AI tsunami on its way. 280 00:25:41.020 --> 00:25:46.181 And there's certainly a lot of interest, a lot of dialogue, a lot of projects, and a lot of work going on here. 281 00:25:46.701 --> 00:25:51.102 But where I think I like to point to is, say, tangible use cases. 282 00:25:51.202 --> 00:26:00.865 One I saw recently was applying AI to account receivable collections and the dunning processes to accelerate collections and reduce manual efforts 283 00:26:01.885 --> 00:26:05.950 Lower days, sales outstanding and improve the cash conversion cycle. 284 00:26:05.990 --> 00:26:11.937 And this demonstrated how clients are looking to embed AI directly into their operations and decision making again. 285 00:26:11.977 --> 00:26:20.566 I mean, looking ahead, I think AI is likely to change connectivity and certainly something we're already starting to see from AI. 286 00:26:21.207 --> 00:26:31.534 a simple exchange between banks and ERPs and TMS platforms, which I think we've covered quite well here, into more sort of intelligent orchestration of workflows. 287 00:26:32.395 --> 00:26:41.702 Rather than treasury teams sort of manually extracting information and then, say, triggering actions, you can see an environment, and we're starting to see them emerge globally. 288 00:26:42.482 --> 00:26:55.966 in a bunch of cases where AI agents could analyze positions, identify exceptions, recommend actions, and ultimately execute approved tasks across connected systems. 289 00:26:56.046 --> 00:27:04.148 And all of those things you mentioned, you know, cash flow forecasting for detection, and liquidity placement, we're seeing work around that space. 290 00:27:04.168 --> 00:27:04.288 So 291 00:27:04.888 --> 00:27:19.876 I think the next step may be sort of agent-based treasury where AI not only forecasts liquidity, detects fraud or prioritise collections, et cetera, but also collaborates across banking payments and ERP ecosystems to sort of drive execution. 292 00:27:20.577 --> 00:27:27.140 I think it's that shift from information into execution that is getting very interesting in this space. 293 00:27:27.180 --> 00:27:33.344 And I sense from the clients that we speak to and the feedback is suggesting very strong interest in technology 294 00:27:34.104 --> 00:27:50.523 in these things being scalable commercial ai and treasury uh use cases not not just proofs of concepts and and they need to deliver some kind of operational efficiency or or or improve the cash outcomes that our clients are you know looking to achieve 295 00:27:51.043 --> 00:27:52.124 Absolutely echo that. 296 00:27:52.164 --> 00:27:55.486 That's the kind of things we're getting questions around on our side as well. 297 00:27:55.546 --> 00:28:01.031 And people looking for case studies on those already, although it's very early days. 298 00:28:01.131 --> 00:28:13.480 But I think, Peter, you kind of mentioned that, you know, it's very exciting times, but we know underneath we need to have a lot of caution around it as well when we're moving towards that execution piece. 299 00:28:13.520 --> 00:28:17.263 So, Amber, let's bring you in here because tell us a little bit about 300 00:28:17.903 --> 00:28:24.446 A, how connectivity enables those use cases that Peter mentioned, prepares us for this agent to agent interaction. 301 00:28:24.766 --> 00:28:41.975 But also once we get there, once we've got agents accessing this data, initiating actions, how do we then need to bring in things like know your agent, permissions, auditability, and of course that human accountability that we speak about so much. 302 00:28:42.485 --> 00:28:43.246 Yeah, sure, Eleanor. 303 00:28:43.286 --> 00:28:51.512 So, I mean, for us, we know bank connectivity is really key for AI because the intelligence is ultimately only as good as the information it can access. 304 00:28:52.072 --> 00:29:01.519 So whether AI is sitting on our client's side or within a bank channel, we need to make sure we can deliver timely and accurate data, the right context, and of course, clear permissions. 305 00:29:02.200 --> 00:29:06.483 And we're already seeing AI change the way that connectivity works with our customers. 306 00:29:06.543 --> 00:29:11.327 So, for example, we've started using Model Context Protocol, or MCP for short, 307 00:29:11.707 --> 00:29:15.269 So we can securely connect our APIs into our client's AI tools. 308 00:29:16.170 --> 00:29:21.813 This builds on trusted security and permissions that we already have in place for clients using APIs. 309 00:29:22.134 --> 00:29:28.197 So it's a really nice way to sort of start introducing that to customers in a secure way that they're familiar with. 310 00:29:28.918 --> 00:29:35.382 And we're also working on how we can let users use their HSBC net credentials, for example, to log on 311 00:29:35.862 --> 00:29:38.904 using our MCP into their AI tools like Claude. 312 00:29:39.404 --> 00:29:49.891 And look, these approaches work well when there's a person in the loop, but as we move more toward those AI agents taking more action independently, we're going to need to go a step further. 313 00:29:49.911 --> 00:30:00.357 So that's where things like Know Your Agent and being able to identify the agent, their permissions, the way we think about oversight and accountability definitely needs to evolve. 314 00:30:00.847 --> 00:30:02.947 Well, go on, tell us a little bit more about that, Amber. 315 00:30:02.967 --> 00:30:04.988 What do you see in terms of that evolution? 316 00:30:05.328 --> 00:30:09.108 Yeah, so I think if you, you know, we think about it, you've got humans doing something today. 317 00:30:09.228 --> 00:30:12.789 In the future, AI agents will take on more parts of that workflow. 318 00:30:13.269 --> 00:30:15.169 And so I think it keeps simple, right? 319 00:30:15.229 --> 00:30:19.770 It's really every decision must remain clear, controlled, and audible. 320 00:30:19.970 --> 00:30:25.051 And a good way to think about it is it's no different to whether that action was being taken by a human. 321 00:30:25.551 --> 00:30:29.852 And so this means we need to go beyond just testing AI models when we deploy them. 322 00:30:30.472 --> 00:30:38.095 We need to continue to have oversight on how that AI is performing every day because LLMs change, the data underneath changes. 323 00:30:38.215 --> 00:30:52.122 And so we need to think about what information is being used, where is it being sourced from, which model and version are involved, and what thresholds do you need to put in place to make sure you have human review or approval where it needs to happen. 324 00:30:52.962 --> 00:30:56.284 And that's where we always talk about accountability needing to remain clear. 325 00:30:56.304 --> 00:31:06.830 And I said that before, you know, AI definitely helps people reach faster decisions, better informed decisions even, because you can process really large amounts of information quickly. 326 00:31:06.870 --> 00:31:07.650 And that's great. 327 00:31:07.790 --> 00:31:19.317 But it is important that we don't become complacent and we continue to look at what is the AI producing in terms of decisions and definitely make sure there's a human in the loop for higher risk outcomes. 328 00:31:19.877 --> 00:31:23.019 And the permissions of the AI agents are an important part of that. 329 00:31:23.059 --> 00:31:36.167 So again, building on the know your agent, it's really about defining who the AI agent is, what data they can access, what actions they're allowed to take, and then how do we control what funds they have access to as well. 330 00:31:36.788 --> 00:31:38.729 Good news, the industry is responding. 331 00:31:38.749 --> 00:31:41.050 And even this morning, I saw another announcement. 332 00:31:41.290 --> 00:31:47.014 There's a newly formed Agentic Payments Alliance that's come together with Visa, MasterCard and Fireserve. 333 00:31:47.574 --> 00:32:01.178 I'd say they're definitely starting to look at more the e-commerce space and consumer and business payments, but there's no doubt those standards that they work on for agent identity and authorization are going to slip into more of the treasury B2B space as well. 334 00:32:01.838 --> 00:32:07.380 And for us at HSBC, we're working really closely with our new Singapore AI Innovation Centre on this with... 335 00:32:07.740 --> 00:32:10.922 MIS as well on what the right approach is. 336 00:32:10.942 --> 00:32:25.391 So as Peter said, as you said, the opportunity is significant, but we need to make sure we have the right controls alongside the technology so we can all operate safely and make sure we can rely on the agents to be accurate. 337 00:32:25.851 --> 00:32:26.572 Absolutely. 338 00:32:26.732 --> 00:32:32.016 And I liked what you said about the complacency from people as well, Amber, because that is so, so important. 339 00:32:32.056 --> 00:32:47.367 And there's some interesting studies out there looking at how humans behave when they're getting outputs from AI agents and how many mistakes they catch at the beginning of the task and how many mistakes they catch after a few hours of the task and it all changing. 340 00:32:47.407 --> 00:32:49.649 So there's so much for us to think about. 341 00:32:49.749 --> 00:32:53.672 But yeah, lots of exciting opportunities coming down the line as well. 342 00:32:54.212 --> 00:32:57.236 Peter, I'm going to give you the tough task of wrapping up for us. 343 00:32:57.296 --> 00:33:00.200 I know we've covered absolutely loads today. 344 00:33:00.320 --> 00:33:05.506 And we've looked at organisations having very different levels of technical readiness. 345 00:33:05.546 --> 00:33:09.431 They also have different levels of investment capacity, risk appetite. 346 00:33:09.551 --> 00:33:09.772 So, 347 00:33:10.272 --> 00:33:25.510 For our audience, for our listeners out there, what is a realistic first step, would you say, Peter, for a treasury team that wants to improve their connectivity, but not necessarily going out on some costly big bang transformation? 348 00:33:26.111 --> 00:33:29.455 And tell us a little bit about how the bank can help there as well. 349 00:33:29.847 --> 00:33:45.130 Well, firstly, what I would say is I feel old in this conversation because I remember the days when we were delivering payments via fax machines and moving to offline versions of banking on a computer into online host to host. 350 00:33:46.211 --> 00:33:54.156 you know, setting up the first closed user groups with Swift now into, you know, more advanced versions of host to host, you know, TCP IP protocols. 351 00:33:54.216 --> 00:34:02.101 And then, you know, where we are in API land and, you know, listening to Amber speak about model context protocol layers for agents to talk to agents. 352 00:34:02.201 --> 00:34:04.903 What an exciting time to be in the payments business. 353 00:34:04.943 --> 00:34:10.927 But, you know, to your question, I think the realistic path forward is evolutionary, not revolutionary. 354 00:34:11.528 --> 00:34:25.949 Over the next three to five years, most treasuries will still operate a hybrid model of some combination of existing channels, such as Swift or host-to-host and banking portals with targeted option of APIs. 355 00:34:26.627 --> 00:34:31.448 embedded banking and eventually, you know, some AI enabled interactions. 356 00:34:31.668 --> 00:34:33.589 I mean, as Amber pointed out. 357 00:34:34.009 --> 00:34:40.790 So I think the goal is not to replace everything at once, but connect the right capability to the right treasury outcome. 358 00:34:41.230 --> 00:34:48.372 You know, for treasury looking to improve connectivity without a costly transformation program, the first step is going to be to identify 359 00:34:49.112 --> 00:35:01.917 you know, a handful maybe, or a few, perhaps even a single high value use case, such as real-time cash visibility, or payment status tracking, or automated reconciliation, and solve that problem first. 360 00:35:01.957 --> 00:35:04.498 Sort of start small, prove the value, then scale. 361 00:35:05.038 --> 00:35:08.661 You know, the bank's role is to co-design a roadmap that grows with the client. 362 00:35:09.321 --> 00:35:11.363 And there is a basket of tools available. 363 00:35:11.783 --> 00:35:17.568 We don't offer fax machine payments anymore, but, you know, definitely the more modern versions of the tools are there. 364 00:35:18.148 --> 00:35:30.418 Now, that means, like, using proven interactions, you know, sandbox testing and, you know, standardized connectivity patterns or clients often ask for reusable architectures that minimize implementation efforts. 365 00:35:31.459 --> 00:35:33.500 Amber's touched on our approach to that. 366 00:35:34.101 --> 00:35:43.608 All of this so that they can create a foundation for future capabilities, such as AI agents, which are most certainly on their way and already here. 367 00:35:45.029 --> 00:35:49.713 These embedded workflows, which takes away a lot of the implementation setup. 368 00:35:50.455 --> 00:35:56.097 So really connectivity becomes a journey of, let's say, incremental value rather than a big bang project. 369 00:35:56.377 --> 00:35:57.418 Yeah, absolutely. 370 00:35:57.558 --> 00:35:58.578 That's the best way. 371 00:35:58.678 --> 00:35:59.579 But Peter, thank you. 372 00:35:59.599 --> 00:36:01.239 You gave me some nice memories there. 373 00:36:01.399 --> 00:36:09.543 My dad used to have a home office with a fax machine and I was forever tasked with changing the, you know, the thermal paper and it would go black. 374 00:36:09.563 --> 00:36:09.883 Yeah. 375 00:36:10.303 --> 00:36:11.844 So you gave me a lovely memory there. 376 00:36:12.024 --> 00:36:13.864 I forgot to mention the CD-ROM. 377 00:36:14.124 --> 00:36:15.065 I'm sure there are others. 378 00:36:15.565 --> 00:36:16.365 The floppy disk. 379 00:36:16.985 --> 00:36:21.066 The City Run was still around not that long ago. 380 00:36:21.186 --> 00:36:21.626 It was. 381 00:36:21.927 --> 00:36:22.687 Oh, my goodness. 382 00:36:23.267 --> 00:36:24.167 Those were the days. 383 00:36:24.247 --> 00:36:28.248 But it's amazing how quickly things are moving on. 384 00:36:28.348 --> 00:36:35.650 And it's great to hear from both of you about what you're seeing in terms of best practices and completely agree with you that, Peter, it's... 385 00:36:36.170 --> 00:36:39.192 is not about necessarily this big bang transformation. 386 00:36:39.252 --> 00:36:44.695 You can do it incrementally, start with that outcome that you want to improve, choose something really tangible. 387 00:36:45.055 --> 00:36:48.877 Then you can look for the right connectivity for the job and build from there. 388 00:36:48.917 --> 00:36:51.639 So thank you both ever so much for coming on. 389 00:36:51.679 --> 00:36:53.620 It's been a real pleasure to have you here. 390 00:36:54.000 --> 00:36:59.864 And for those of you listening, just a reminder that this is part of TMI's special Redefining Treasury series. 391 00:37:00.324 --> 00:37:02.747 produced in collaboration with HSBC. 392 00:37:02.927 --> 00:37:04.790 And I hope you've enjoyed this discussion. 393 00:37:04.870 --> 00:37:06.572 We've certainly enjoyed recording it. 394 00:37:06.752 --> 00:37:10.537 And please do check out the other episodes in the series. 395 00:37:11.078 --> 00:37:11.899 Thank you. 396 00:37:38.152 --> 00:37:41.157 Thank you for joining us at HSBC Global Viewpoint. 397 00:37:41.598 --> 00:37:42.981 We hope you enjoyed the discussion. 398 00:37:43.742 --> 00:37:46.427 Make sure you're subscribed to stay up to date with new episodes.