To kick off the episode: Remote work was supposed to make jobs more accessible — but for entry-level workers, it might be doing the opposite. New research suggests remote roles are squeezing junior hiring more than AI (for now). Then: why Gen Z has turned sardines into the snack of the summer — and a surprisingly lucrative business. Plus, the Big Picture: Ottawa’s fighter jet rethink, a new BC Conservative leader, Alberta wildfires, and more.
And we end off with an exclusive interview with Chad Koziel, Associate Vice President AI Product at TD. Check out our notes below:
The Peak: Hey, Peak pals, we are joined today with Chad Koziel, who's the associate vice president of Layer Six Gen AI delivery at TD. And we're going to have a really interesting discussion today. There's a lot of talk about Gen AI and how transformative it's going to be, but not a lot of discussion around the actual practical applications of that at the organization level. So that's why I'm really excited to jump in with you today, Chad. Do you want to just take a second and introduce yourself?
Chad: Thanks and it's great to be on here, Brett. So I'm Chad Koziel. I build AI. That's what I do. I make it real. I put it in clients' hands and colleagues' hands. And my job is to do so in a way that they love, in a way that they want to engage with, they want to adopt, they want to play with and that ultimately drives a positive impact for them.
The Peak: That's awesome. And I imagine a lot of this is happening behind the scenes. So how is AI working at TD right now without clients even knowing that it's there?
Chad: So, if you're a TD client, you can expect today that if you call in with an issue, you get helped better, you get helped faster than ever before. You may or may not notice it on a call-by-call basis, but it's there. In the background and across TD, AI capabilities are being deployed to make everything simpler, to make everything faster and specifically to get information to people faster. So that concrete example I gave you, in the background, there's a tool, a virtual assistant that allows a frontline colleague to solve an extremely complex customer query in seconds, empowering them to help you in that moment. Whereas in the past, they might have put you on hold while they looked stuff up. They might instead have put you on hold and contacted one of their colleagues who could help them work through it before coming back to you. So we're seeing a 12% reduction in those types of things. And it's not just in our contact center. So we have these fit-to-purpose, fast, accurate virtual assistants deployed in TD wealth, in branches, in insurance and TD securities. All of these driving, perhaps imperceptible, perhaps perceptible, but nevertheless impactful improvements to the client experience.
The Peak: That's a really great case study of how generative AI is helping TD clients and as a TD client myself, that's really great to hear. There's a lot of excitement about this technology, but there are some fears as well about it. You know, Gen AI is here and it's brought a lot of concerns with it about how our lives and jobs are going to change. What do you tell someone who's more concerned than excited about the impact of generative AI?
Chad: What I described previously in contact centers is the tip of the iceberg. There's a lot that I have the privilege of having a front row seat to that I can't share publicly, but the impact is enormous, it's profound. And that's leading to companies and people in every sector, in every field to adopt AI, generative AI to improve the quality, speed and availability of service. So maybe, you know, in this conversation with this concerned person, I would then say, seek it out, play, toy with generative AI. Just as an example, like I've used it to dramatically change my reading habits, like in volume and complexity by feeding it things I liked, things I didn't and using Gen AI to take me off the beaten path into areas I may not have previously read from. Authors who are maybe terrific but didn't sell, geographies outside of a western-centric sort of award circuit. And that's a toy example, right? But this is the kind of thing somebody can do to gain some familiarity with it, maybe even comfort. At TD, we put trust at the center of our relationship with clients. And when we move forward with AI, we will do so responsibly, transparently, reliably, safely, all of those things. That's how I try to bring some of this comfort to somebody maybe a little concerned about generative AI.
The Peak: That's very helpful. And you mentioned it at the end there, but I would love to just dig into how TD is actually keeping AI safe and secure when handling clients' money and personal information.
Chad: That's a great question. First is we don't use AI to handle your money. So AI is probabilistic, it's guessing things and we don't guess where your money should go and frankly no algorithm at TD should do that. You decide where it goes. We might use AI, we will use AI to make it easier for you to do what you want to do with your money. And we should make it a lot easier for you. We will never do it for you. Now, the personal information angle though is pretty interesting. When TD's AI consumes your information, it's in a cloud vault in an area fully and completely controlled by TD that's not accessible to any of our outside partners. And that's not going to change. We are going to protect that information and the handling of that information in a way that preserves your trust. And related to that, like when we do go and use AI, we're going to try to build that in a way that's remarkably human. So as an example, ensuring that empathy, intuition, care are at the heart and frankly quantitatively evaluated so that we know, not just we think, we know that the experience we're giving you with AI is one that you can trust is doing that in a way that feels right to you. The last key point because it's an area of particular focus and one where we have a bunch of research is around trustworthy AI. So TD, we have an entire team of scientists dedicated to evaluating for privacy, security, fairness, accountability, explainability, and not just ensuring that those are all built into the models, to the AI that we release, but also to ensure that that is continuously monitored so that those AIs do not stray over time, giving us the confidence to put them into market and have them be in the hands of our customers and colleagues.
The Peak: That's really great to hear and very comforting. Now, we know that AI is going to change banking for Canadians, but what parts of the experience will always need a human touch even as this technology evolves so quickly?
Chad: So, a concrete example, let's call it the estates process. So you've just had a terrible loss in your family. And not only do you have a terrible loss, you're dealing with an incredibly complex situation, financially, legally, emotionally. I guarantee you you want somebody sitting across from you helping you through that. But what you want is for that person to be able to work through this quickly with you to not take days or weeks to sort through hundreds of pages of documents. You want them to be able to do that in minutes to give you that trusted advice right away and with their human touch, with empathy, with intuition. That's an example, but we can think of a bunch of similar sort of challenging events, right? Whether it's job loss, retirement planning, thinking of buying a home. You're always going to want a person there and what we feel is that that person will, however, need to be empowered by AI to do that in the way that you expect.
The Peak: Yeah, that makes a lot of sense, especially as we think to what this next stage of AI is going to look like. And actually on that note, just three years ago, we entered a new phase of AI when Canadians got access to Gen AI chatbots at home. What's coming next for AI and how is it going to change banking?
Chad: Great question. So today, tomorrow, right now, the buzzword is agentic AI. And so what we're talking about are generative AI capabilities that can execute complex tasks with a high degree of autonomy. It is here. It is here, it is driving fundamental change. And it will do so in an obvious way, faster, better than humans were capable of. And at TD, you'll see this in the real estate secured lending process or getting a mortgage, right? We are inserting agentic AI at every step to make that process faster and better for our clients. And specifically what used to take hours, days is now seconds and minutes. And that creates an enormous opportunity for cost savings, for acceleration. And you'll see that happen across TD, not just in real estate secured lending. Here's what I hope we sort of see next, which is intuition, nuance and teasing out deep understanding. So in playing with generative AI models, in seeing what they're capable of, they can find these patterns. For example, just a few words that indicate maybe you're a little bit hesitant about your current debt situation as you pursue a new mortgage. Yeah, we could just offer you a mortgage. Sure, you're pre-approved. Great, right? But is there something we could do here that would help you out a little more? Could we help you consolidate some of your debt as part of this? Could we help you feel more comfortable with the ends and outs of mortgages and HELOCs as part of it? Spend that time with you, improve not just your financial well-being, but your confidence. I think what we're going to see is generative AI empowering that type of intuitive, emotionally resident experience in a way that we've never seen from AI in the past. And so that's what I think and that's what I hope is next.
The Peak: That's very exciting and I guess we'll have to see where it goes and how quickly we get there because it seems almost inevitable that we will get there. But Chad, thank you so much for joining us today. It was just great to dig into the practical applications of AI, especially around safety. I think that was really comforting for a lot of folks to know. So thank you again, Chad. It was great chatting with you.
Chad: Thank you very much, Brett. I sincerely appreciate your time today.