DerivSource Podcast Transcript (Listen Now)
Julia Schieffer:
Hello, and welcome to this DerivSource podcast. I’m Julia Schieffer, the founder and editor of DerivSource.com. You’re listening to episode two of our Post-Trade Digitalisation Podcast series, sponsored by Torstone Technology. In the first episode of the series, we explored the top trends influencing post-trade digitalisation both from the view of addressing processing, but also supporting new market developments, such as the rise of digital assets.
Naturally, we have already touched on the role of emerging technologies somewhat in episode one, but in this episode we are taking a deeper and more practical look at how firms can use newer technologies, such as cloud computing and AI, as part of their digitalisation post-trade plans today, and also in the long run.
And with me, I have two industry experts to discuss this. Peter Kennedy, Partner at Capco, a global management and technology consultancy dedicated to the financial services and energy industries, and Amrik Thethi, Chief Technology Officer at Torstone Technology, a SaaS platform for post-trade securities and derivatives processing technology. Welcome to you both.
Can you both introduce yourself to our audience first? Starting with you Amrik.
Amrik Thethi:
Hi, I’m the CTO of Torstone Technology. That’s chief technology officer. So I look after the technology direction, what we’re looking at now, the next six months, next year, next five years, and looking for industry trends and trying to create technologies that help people solve their problems.
Julia Schieffer:
And Peter?
Peter Kennedy:
Hi, I’m Peter Kennedy. I’ve been working in capital markets since the early 90s – initially as a software developer and subsequently I was a project manager and architect. And I’m now head of the UK technology practice, at the consulting company, Capco.
Julia Schieffer:
Let’s begin by getting right to what the most common ‘newer’ technologies firms really use today. Peter, what technologies do you see being used by firms as part of their digitalisation strategies for the post-trade operations when you work with these clients? What are they actually using?
Peter Kennedy:
Yes, the most common ones really are cloud-based services. Cloud itself isn’t new, but new capabilities on cloud emerge all the time. And to a large extent as an industry, we’re only just starting to explore the possibilities of what we could do with the inherent capabilities of cloud. We’re also seeing some level of process automation as well, and an increase in the whole area of process automation.
A lot of that is actually driven by some elements of what’s termed ‘process mining’, which is essentially a way of, to some extent, automating the process of figuring out the data flows and the processes within an organisation. And that’s especially relevant to the whole post-trade area – that’s a notoriously complex problem to solve.
And then finally we’re seeing clients upgrade or replace their legacy platforms with either a microservices type architectures or some form of externally provided and generally cloud-based services.
Amrik Thethi:
From my perspective as a vendor, we are the ones providing those services. So we provide a cloud-based platform. So what we do is replace legacy platforms with a cloud-based solution that is upgraded automatically and regularly for the customers. But at the same time, we are seeing them use other technologies with our systems such as RPA, robotic process automation, and sometimes not so much AI, but certainly people are experimenting with it. I think that’s more for the larger organisations, but certainly scripted automation and things like that, we’re seeing quite a lot of.
Julia Schieffer:
On the topic of cloud, this is not a new technology, and it’s one that many of us use on a personal level daily, yet my impression is that cloud isn’t as prevalent in post-trade processing today amongst firms as it could be. Amrik, what is your view on the use of cloud technologies and how it could be used more practically for post-trade operations?
Amrik Thethi:
Yes, definitely. Cloud is the big one. As you said, it’s not new technology, but the post-trade space has historically been under invested in terms of change, and they’re catching up now. So we’re seeing a lot of interest in our products and other vendors’ products. And the question used to be asked four years ago, ‘what is cloud?’ ‘Why should we use it?’ Now the question is, ‘why are you not in the cloud for any vendor?’
We were a bit ahead of the game, but other people are on to it now. And the customers are finally caught up and it’s because the benefits are clear to everybody. And more importantly, regulatory hurdles and other question marks have been resolved. I think the reliability and also transparency and access to data and things like that. So it’s all guns blasing as far as cloud is concerned.
But linked to that also is the API story in that systems were traditionally very closed, files being sent and files coming out to have a much more in-depth and flexible integration. Having APIs, application programming interface, ways that you can call into the application from another application and get it to do things has become the de facto expectation now in your product. Every vendor’s product has some level of ability to call into and do things, which allows you to create systems that don’t necessarily exist in independence. You can take a bit of this, a bit of that and put them together and create a bigger system and tie that together with things like RPA and other things, and perhaps even AI, you can create sort of a much more sophisticated post-trade environment.
Julia Schieffer:
Beyond cloud, what about other newer technologies such as AI or artificial intelligence? In your experiences are firms actually adopting this AI technology and adopting it to their post-trade processes? Peter, what do you see?
Peter Kennedy:
Yes, they are adopting AI. Now there was an initial wave of robotic process automation, and that’s been around for a while. I’m never quite sure whether that really counts as artificial intelligence, but we are certainly seeing the capabilities of AI get smarter in terms of its ability to deal with natural language processing, and its ability to deal with pattern recognition and much more complex decision-making.
I once worked at a client where if you sat in the trading floor after a certain time of the day when everybody had gone, it was amasing how much voice activity you could still hear going on the desk, almost all of which was actually automated. So yes, the post-trade world is I think increasingly seeing all sorts of different elements of AI and not necessarily just process automation.
Amrik Thethi:
Certainly when it comes to AI, such a wide-ranging term, probably a bit abused in that people will try to put the AI term on anything even when it’s pretty much a standard process. I prefer the term machine learning, and it’s certainly helpful for situations where you’ve got a huge amount of data and you don’t know what to do with it, and you try and look for patterns in it. It has helped with things like reconciliations in the back office some large organisations where ideally they shouldn’t have to reconcile all these things and they shouldn’t have all these breaks in the first place, but they do and fixing everything in one go is going to be a nightmare so they apply some machine learning to look for patterns and at a depth a human couldn’t do. Ideally someone would go and sit down and create rules for everything, but it’s too big for probably a human to do, but a machine can crawl through that data and find the patterns and then encode those as rules and then help the reconciliation process. Those are the kind of concrete examples. I think we’ll see more and more of those things, very specific use cases that come up when you have large amounts of unstructured data and a machine that could spend hours and hours just crawling through it.
Julia Schieffer:
Fantastic points. So we’ve touched on the specific technologies already, but taking a step back, let’s talk about the drivers. What are the more typical drivers behind a firm deciding to take the plunge, to move beyond their legacy systems and adopt some of these newer technologies as part of a digitalisation strategy?
Amrik Thethi:
There’s usually a number of drivers, but as our salespeople always say, there has to be one compelling event that drives an organisation to start changing their post-trade systems. Those compelling events vary, but they are usually to do with something, external influence such as a merger or a new business that’s being taken on, a new opportunity to make money some other way, and you’ll find that the legacy system that’s been there for 10 years, maybe 15 years, isn’t capable of doing that function or they just got three systems that really could be one system.
And so that’s usually the compelling event, some external business reason for doing it, but then they usually take the opportunity to look at what other things can be cleaned up and how they improve things and not replace like or like but replace something with something that’s categorically bigger and better, and more in line with the future. And systems in the post-trade world tend to live for a very long time, so the decisions around them tend to be quite carefully made.
Peter Kennedy:
That’s obviously right. So all of our clients have some kind of problem with legacy systems, whether that’s controls or capacity, or operational risk, or unsupported software. And I’m sure we’re all quite familiar with the challenges of working with older technology.
I think the key point for me here is that we’ve seen a shift in technology just over the last five years, and such that it’s becoming significantly cheaper, quicker, and easier to adopt new and often custom technologies, and that’s a combination of all sorts of things to do with cloud and containerisation. But also the fact that we, as an industry are getting better at high quality engineering.
And I think that’s tipped the balance towards newer technologies. It’s just easier than it ever has been before to create new technology. Now that’s not to say it’s easy. These things are never easy, and there are lots of good examples of failures too, but fundamentally the driver is an economic one.
Julia Schieffer:
And on the topic of drivers, in the previous podcast episode we had discussed the ongoing pandemic, being a catalyst for digitalisation generally. Amrik, in your view, has the pandemic had any impact on digitalisation trends?
Amrik Thethi:
Our customers had no problem with the pandemic because they were using our system, which is cloud-based and available everywhere. But that’s just me. We are definitely seeing that this is an issue because a couple of things… if the business processes they have require people to be shuffling paper between themselves, that’s a bit of a problem when they’re all at home.
So the first thing is digitise your processes so that you don’t have to be shuffling pieces of paper and signing authorisations and things like that. And people forget that these things are happening, but you only realise it when the person is no longer available. Simple things like that are what this has really helped with. So yes, I think it is definitely concentrating people’s minds on what can be improved.
Peter Kennedy:
I think in the short term, the clients that I’ve seen have responded extremely well in the post-trade space to the pandemic. And I think short term it’s shown the measures that they had in place to be able to work remotely, and to be able to deal with unforeseen events have worked quite well. I think, however, in the medium to long term, it has really focused minds on the areas which are still too manual and still too reliant on real point skills and key person dependencies.
My clients did have to have some teams in the post-trade space (middle and back offices) who simply had to be in the building and had to be able to interact with each other, and that’s been possible in this pandemic, but in other eventualities may not be. So I think longer term, we probably will be seeing architectural changes and design changes, and technology driven by those concerns even though actually in the short term it’s responded very well.
Amrik Thethi:
I definitely agree with that. You can’t predict everything, and you have to live the problem before you find the solutions.
Julia Schieffer:
We’ve talked about drivers behind the change, but are there any reasons why a firm would choose to stay with legacy systems or older technologies? Peter, what’s your view?
Peter Kennedy:
Fundamentally, it’s easy to dismiss old technology just because it’s old and to admire the cleverness of cool new digital technology. But legacy systems often do the job, and they often do that in a highly cost effective way. So there may be challenges with legacy technology like out of date hardware and unsupported operating systems and so on which oblige you to change it.
But sometimes none of those apply or they’re outweighed by the cost and risk of new technology. And in particular, in post-trade we’ve seen lots of good examples of that. Some of the core systems and vendor packages and services which support the post-trade world have been around for a long time and continue to be very effective.
Some banking core operations technology are into their fifth decade of operations, and they’re still not looking likely to get replaced anytime soon. So it isn’t necessarily the case that new technology is inherently better than old technology. And in some cases, the old systems continue to do a very cost-effective job.
Amrik Thethi:
Yes, I definitely agree that you shouldn’t change things just for change’s sake, there has to be a good reason for it. So my definition of legacy system is one that you are scared to change. Are you scared to update it because it’ll break or it will do something unexpected? If you can change it and you can support it, and it’s not running on something steam powered, then you’re probably okay with it.
It only becomes legacy as soon as it doesn’t support the business that you’re trying to run on it. If it’s a mainframe and the software that runs on it is produced by a vendor that supports it, then that’s not a legacy system, that’s just a system.
Julia Schieffer:
I think that’s a very good distinction. Thank you Amrik. Looking at the practicalities of any large or small digital transformation plan, what are some of the common challenges that firms might encounter when moving to newer technologies? Peter, what do you see amongst your clients and in your own experiences?
Peter Kennedy:
There are a number of challenges, and they aren’t necessarily technology related challenges. I think first of all, the business angle of the post-trade world remains quite complex, and legacy systems actually hide quite a lot of that complexity. So hooking up to the world’s exchanges and clearinghouses, custodians and so on, is just harder than it sounds. And the underlying financial products that the post-trade world deals with are complex and rely on a lot of data to make them make sense and to make them work.
New technologies are not a panacea and they need to be implemented with care. So cool architectures like microservices are suitable for some applications, but not necessarily for all. Selecting the right architecture patterns is really the first thing to get right. And then accelerators like low code environments can go wrong. Systems still will need to be built and implemented in line with standards and principles. And in a world of autonomous agile teams that needs some careful thought and really clear communication to get right.
And then the third point is culture. I think working in agile like ways is not necessarily that easy, it needs complete commitment across the whole organisation, and that’s not always easy to do. And then finally, I think from a skills point of view, finding the right skills to implement a really successful digitalisation strategy. Those skills are at a premium and financial services has to compete for those skills against many other industries in a way that it’s actually much harder now than it was 10 years ago. You can do a lot with technology, but the code still needs to be right. So those engineering skills still need to be right.
Amrik Thethi:
Some really good points there. I think the first thing people have to understand is, what are they trying to achieve with a digitalisation strategy? They’ve got to really have a good idea of their end goal. The other point that is that the complexity there is real and it’s not caused by the system, is caused by the whole environment. The post-trade space involves a number of systems, the market infrastructure providers such as CSDs, exchanges, custodians, regulators, and they’re all at different levels of sophistication. It’s no point you saying, are you on a real-time settlement system when the custodian or the CSD only accepts files once a day, obviously you can’t change the world in one go.
And people have tried that, but it doesn’t work. The whole industry has to move together, and that complexity is never going to go away. Really what you’ve got to do is work out what bits you want to benefit from for your customers and improve those bits and have a clear idea of it.
Julia Schieffer:
One theme that always comes up with looking at both addressing inefficiencies of post-trade processes and implementing newer technologies is the problem with data. Let’s talk about this. What are the challenges that both of you experience with data? Peter, with you first.
Peter Kennedy:
Sure. So at Capco, we’re doing a lot of work on data. So data governance and data lineage. And over the past 10 years or so, the focus I guess, has been primarily on the output, so the regulatory reports and what goes on externally. What we’re seeing at the moment is a huge focus on internal data lineage with the very specific aim of enabling a subsequent wave of artificial intelligence. Fundamentally machine learning and AI is only really as good as the data that supports it. That’s the explicit goal of an awful lot of the data work that we’re doing at the moment.
Amrik Thethi:
Yes, I mean, that’s something we find as the old adage goes, ‘garbage in, garbage out’. And if you can’t trust your data, everything else is almost irrelevant. As we do transformation of companies, by far the biggest problem when we’re doing that is if the quality of the data that they currently have; it tends to be spread across a bunch of disparate systems or spreadsheets, or it tends to have lots of duplications and unnecessary stuff that’s been there for ages and no one can explain. So, data is definitely something that people need to get a hold of, and it is something that I think has been recognised, and people do have programs to improve it.
Julia Schieffer:
Given the challenges discussed, Amrik, do you have any advice that you would give a firm who is about to embark on a digital transformation project? What are some of the critical steps that they really need to do or make in order to get things done right?
Amrik Thethi:
We’re a post-trade vendor so we go in and replace existing systems. We’re doing it all the time. The customer is doing a very, very infrequently; when we go somewhere, it’s usually the first time they’ve done it in 10 years, 15 years.
So the recommendations are simple things like, have a very clear idea of what you’re trying to achieve. Don’t try to achieve too much in one go, do it in small chunks. Sort of following the agile approach. Try to split the problem up into smaller chunks that makes sense for your business and then prove one before you move on to the other one.
Hire specialists who could help you out. One of the issues we have is that sometimes people try to wear two hats where they’re running their day job, and they’re trying to implement a change programme, and they’re trying to think of things five years ahead.
The main takeaway is do it in small chunks, prove those things work, and then move on to the next one. It’s not always possible in a post-trade platform because it’s right in the center and the heart of your business and you can’t have two systems necessarily running at the same time. But if you go divide your business by say asset class or by maybe the business line, do one of them first prove that, show the business benefits and take it to the next level.
The main thing is you might have an extended programme that goes on for say a year, as opposed to a really compressed program that tries to push everything into a six month horizon where you try to change everything in one go.
Julia Schieffer:
Final question for both of you, we are in the midst of a very changeable time as a result of this pandemic. So of course it’s harder than ever to predict the future, but do you both have any views on if digitalisation will help firms future proof their operations and particularly through the application of these newer technologies? Amrik, why don’t you go first?
Amrik Thethi:
Yes, I think definitely they will allow firms to be more agile. The point is that having a good agile system allows you to then deal with future shocks to your business by putting it in the cloud, but turning it into something that’s a bunch of components that can be replaced individually by having clear communication approaches means that you can chop and change.
You may find that that bit that deals with a particular function of your business is no longer fit for purpose, but you can rip it out and replace it without replacing the whole system. What would have involved a wholesale change, buying a new house as it were, now can be done at a smaller scale, maybe redecorating a room instead.
So that kind of ability makes you better able to cope with the changes that will inevitably happen to your business.
Peter Kennedy:
As Amrik says, it’s really all about the rate of change. Digital technologies allow us to change our organisations quicker. We’ve probably all seen in particular post-trade environments where the release cycles and the ability to introduce new products and markets is really quite a painful process.
And the point with digital technologies is that we’re aiming from a technology point of view, we’re aiming at fearless change. I think that’s a good way of looking at it. And really what that means is that you can introduce the new products and new customer propositions, and so on, much more quickly, and you can test them and you can iterate them.
And I think moving in the near future, we’re going to see a real disruption in our markets. And the kind of things that we’ve been talking about today are what will allow the financial services industry and in particular, the post trade part of it, to deal with that kind of pace of change to be able to bring those products to customers as well as manage all of that risk and financial mayhem that otherwise might result.
Amrik Thethi:
To put it simply I would say, change the bank should be same as run the bank. It’s an organisational problem. Be ready for things being different, and don’t make it into a big thing, make it into something that you do every day. When you do it every day, it’s familiar and you’re not scared of it.
Julia Schieffer:
The application of these newer technologies is something we are all becoming more familiar with. Thank you Peter and Amrik for sharing your insight with us today. You’ve just heard episode two of our Post-trade Digitalisation Podcast series. You can find the transcript and other relevant resources on our podcast show notes page found on DerivSource.com. And please check out episode one, if you haven’t already listened to it. Up next, and in our final episode, we explore how digitalisation plans for post-trade processing might work in practice. In the meantime, thank you for listening. Join us next time.