Murray Bender: RBC Investor Services presents insights on the challenges and opportunities facing corporate and institutional investors. In today’s podcast, we’re continuing our discussion on artificial intelligence. I’m pleased to welcome Matthew Clark, Director of Innovation and Acceleration at RBC, sharing his insights on this very timely topic. Welcome, Matthew.
Matthew Clark: Thanks. Happy to be here.
Murray Bender: So to start, Matthew, can you tell us a bit about your role at RBC?
Matthew Clark: Sure. I’m the Director of Innovation and Acceleration within Investor Services IT. So that’s the technology side of our organization. Really, what my role is about is about linking innovation and technology into Investor Services.
So at RBC, we have a vast pool of resources of innovative people who are working on what we like to refer to as Horizon 2 technology. And it’s about looking at that strategy, looking at that technology that’s coming up, and seeing where it actually applies to our business. So that’s where I fit in, is actually taking that lens of all this great technology that’s coming about and where it actually fits into our organization.
Murray Bender: So you mentioned all the resources at RBC. I noticed that a recent industry survey ranks RBC very highly for the maturity of its artificial intelligence. What distinguishes RBC from other financial institutions when it comes to AI?
Matthew Clark: So RBC—and you’re specifically referring to the Evident AI survey. So in that one, we rank number one Canada, number three in the world. And the reason why we’re ranking so highly in AI is because we treat it as a core competency. It’s not something that we’re thinking is coming about in years from now; it’s something that we’re operating on right now.
So the biggest one and probably the most public one that we always talk about is Borealis AI. So that’s a company that we own and operate specifically to move RBC forward in the industry. What they are doing is more of a research arm of AI. So they’re writing papers, there’s doctors specifically working in that space.
And the two main products that they work on, one is in P&CB, so the personal and commercial side of the organization, is NOMI. So if you’re an RBC client, you’re probably familiar with NOMI telling you you’re spending too much money.
And then on the capital market side, it’s Aiden, which is really about how we do trading in capital markets. There’s a whole bunch of new stuff with Aiden that came out this last year that I’m sure is very easy to find.
Really that’s what’s distinguishing RBC, is the fact that we’re treating it as products that we’re releasing to clients already. So it’s something we take very seriously.
Murray Bender: So you mentioned personal and commercial banking and capital markets. What about Investor Services, RBC Investor Services? What are some of the AI projects that you’re currently working on in that particular business? And how do they impact clients?
Matthew Clark: That’s a really good question and something I didn’t address. Specifically, Evident looks at the public side of AI. But internally, we use AI everywhere.
So the biggest one is in anti-money laundering. The anti-money laundering team is really just an AI team at the end of the day. They’re doing a lot of data analytics, predictive analysis, and really understanding how our data flows and where anomalies arise and flag them up for investigation.
On the other side, when it comes to incident management, which is always a boring topic to talk about because nobody wants to talk about incidents. But what we do is we will actually scan our systems and do that same type of anomaly detection, which AI is really good at, to determine issues in our systems before they become business-facing.
So a good example might be some overload on systems happening. And historically, we might have known that that was causing incidents that get tracked in our incident tracker. The AI will flag it up saying, hey, this is an anomaly in our system, and it maps to a very well-known problem that we have. So go and fix it before it becomes an issue.
So those are just two examples where internally, we’re doing a lot with AI. And both of those are impacted in Investor Services. So we use both of those technologies because it’s a broad system that we use. I’m just going to give the name of the one for incident management. It’s called AIOps. Everyone across the organization uses that one. And it really does improve our incident management.
Murray Bender: Which would have a direct impact on clients, too.
Matthew Clark: Exactly.
Murray Bender: So if we can broaden the focus a bit now, based on all of your experience with artificial intelligence, what would you say has most surprised you about this emerging technology?
Matthew Clark: Yeah. So this is kind of like the elephant in the room about AI. Everything I’ve talked about so far has been what a lot of people are calling traditional AI, where you’re using a machine-learning algorithm to predict the future, essentially.
The big elephant in the room is generative AI. So that’s where you have things like ChatGPT, Bing Chat, Bard, Gemini is the new one from Google. This is probably the most surprising revelation of AI we’ve seen. So what they’re doing is they’ve essentially scanned large language models, which is like terabytes of text, same with pictures, videos. There’s a whole bunch of different models of AI that—
Murray Bender: And terabytes is a lot of bytes.
Matthew Clark: Yeah. It’s a lot of data. It’s a level of data where, as a human, you’d never be able to read in your lifetime. But all this data exists on the internet. So it’s pretty accessible for these AIs to grab it.
And I’m pretty sure you’re probably someone who’s gone to ChatGPT or Bing Chat and has typed in some questions and seen it generate you real answers. There’s no human behind the scene. It’s just math. And the answers it’s providing are decent. They’re way, way better than you would expect.
Immediately you get into use cases that you could apply that to, which is not like the traditional type of AI we’ve had before where we’ve needed to provide those data sets. And RBC has tons of data internally and that’s why we’ve been so successful in AI so far. But with generative AI, you’re using data sets that aren’t your own company’s. You’re using these large language models. And the predictive text that’s coming out, you’re saying, like, hey, well, I could use that for this use case over here. I need to generate tons of text.
So a really good example, and it’s pretty safe, is documentation. Documenting your own internal processes can be a time-tedious task because you’re just typing up basic English language text. You could easily get a large language model to format that for you and give you this nice text and ask it different questions about providing it in different ways. And that’s really powerful. And that saves you a lot of time and that productivity gains. We want to see those within the organization.
So that’s probably the most surprising thing in AI that’s come out in the last year. And it’s definitely what everybody’s talking about.
Murray Bender: Indeed. What about the future? Where is all this headed? How is AI going to be evolving—continuing to evolve, I guess?
Matthew Clark: So AI and specifically, again, generative AI, it’s here to stay. So we’ve essentially proved in the last year that this is a powerful technology that has a lot of different use cases in everybody’s day to day, even when it comes from writing emails or doing your own job, getting some pictures for your PowerPoints, generating the PowerPoints themselves. Because they have so much information already baked into them, it’s going to be part of pretty much everything you do.
So, a really good example of that is Microsoft 360. So that’s Microsoft’s kind of—they’re incorporating their own AI into all their products. By doing that, we’re going to start seeing that, hey, this product now has an AI tool associated with it, which you can ask it to generate you some text. And you’re going to see that in all of the tools you use throughout your day to day.
Another good example might be in Excel. It’s going to say, hey, you should use these formulas. And now it’s already had some of those predictive things, but it’s going to be in the next level. It’s going to be doing work that you would have said, like, hey, I was supposed to be doing that myself.
So you’re going to get all these productivity gains across all the products that we have. But when it comes to the main use cases RBC looks at, so the big ones we’re really targeting for RBC is code generation. So reducing the time it takes to develop new features and close that STLC (software testing life cycle) feedback loop because it’s really good at that. And there’s a lot of open-source code out there that has been scanned by things like GitHub Copilot, AWS CodeWhisperer.
And then the other one is around our advice centres. So this is going to be a little bit of a weird one, but our advice centres operate on these policy and procedure libraries. And those are vast documents of text, like rich, rich, thousands of documents that describe every single thing we do at the organization. And searching and finding a specific procedure is time consuming. And building up a context within one of these large language models with your own documentation, which is called retrieval-augmented generation, that’s going to really move us forward in terms of how fast we can answer client queries.
Specifically in Investor services, that’s one of the problems that we’re looking at. So that’s what my team is actually trying to apply to the problem set that our client representatives are doing today. So that’s definitely something we’re probably going to see in the next year.
Murray Bender: So stay tuned.
Matthew Clark: Yeah. So if you’re a client calling the advice centre, the hope is that you’re going to get responses faster.
Murray Bender: Interesting. Thanks for your time, Matthew. We really appreciate it.
Matthew Clark: Thanks for having me. This was a lot of fun.
Murray Bender: For insights on a range of topics relevant to corporate and institutional investors, including our previous podcasts, visit rbcis.com/insights. I’m Murray Bender. Thanks for listening.
This content is provided for general information and does not constitute financial, tax, legal, or accounting advice and should not be relied upon in that regard. Neither RBC Investor Services nor its affiliates accepts any liability for loss or damage arising from use of the information in this podcast.