The Productivity Pivot:
AI’s Role in Business Transformation and Growth
Paul McKinlay
Executive Managing Director and Co-Head, CIBC Innovation Banking
Welcome, everybody. Thanks for being here. We're here to have a conversation this morning about AI. Really at the intersection of AI driving productivity, as well as sort of the adaptation, adoption and cultural change that takes around AI. My name is Paul McKinlay. I lead our Innovation Banking team at CIBC. I'm here with two of my colleagues, Dan Lee, Tech Investment Banking from the Capital Markets Group, as well as Dave Gillespie, EVP of Infrastructure and Architecture, and leads a lot of the things AI inside the bank, as well as two esteemed guests we're honoured to have here today, Mark Shulgun, the Co-Founder of Intrepid Growth Partners and Dan Debow, Serial Entrepreneur, Investor, Former VP at Shopify, been around the tech sector for a long time. Mark, I'll move to you quickly. You guys have built Intrepid. Maybe before we do that, a quick thirty seconds on Intrepid Growth Partners, what you guys are up to. A lot of it is around the thesis of AI and productivity. So, we'd love to maybe expand on that.
Mark Shulgan
Co-founder, Intrepid Growth Partners
Yeah, absolutely. So, we set up Intrepid, a couple of years ago. We set up Intrepid with a focus specifically on machine intelligence, because this is the most transformational thing that we'll have ever experienced. And we really believe right now we're still in the very early innings of how this is going to evolve. And so we thought with a, with a purpose built fund that focuses exclusively on machine intelligence, we're just going to increase the odds that we're going to catch the the best companies that are riding through this kind of evolution.
Paul McKinlay
Executive Managing Director and Co-Head, CIBC Innovation Banking
So, one of the things I've heard you and your colleagues talk about, though, is when you make an investment or you're sizing up an investment, really a laser focus on the productivity gains. ... that AI ... so not necessarily just like, is this stack, is this technology interesting, but what is the output or the really impactful productivity gain for that company or for society, or what have you? How do you guys think about that? How do you measure it and how did you come to that thesis?
Mark Shulgan
Co-founder, Intrepid Growth Partners
Yeah, well, we've coined the term ‘ROAI’, to play on return on investment. But it really comes down to that. I mean, we, in every investment that we make, we want to very clearly define what is the return that you're getting from AI. And that's simply, what did the process look like before and what does it look like after. And then using a very detailed approach, what is the uplift that we're getting in terms of what the actual return on investment is? And that really sits at the foundation of everything that we're doing.
Paul McKinlay
Executive Managing Director and Co-Head, CIBC Innovation Banking
So, there's a theme there I want to come back to, but I'll go to Dan for a second around the capital markets funding. We talked about, you know, one of the stats I'd heard this morning is that some of the CapEx and, and spend around AI infrastructure is going to make up a full percentage point of GDP in the US, says the US Treasury. What do you see in the markets from funding valuations perspective, as well as like how it changes the dynamic for both software and AI businesses, but also maybe the rest of the economy?
Dan Lee
Managing Director, Technology and Innovation, Investment Banking, CIBC Capital Markets
Yeah, absolutely. So, before there was any even real adoption at the enterprise, almost $600 billion had been invested into the category, whether it's through CapEx and building data centres, or the money that was raised by the frontier models. Now that we're seeing, real enterprise adoption at scale, where projects are moving from pilots into real, production situations, that funding level's just taken off completely. I’m going to park-aside the CapEx discussion because, I mean, those numbers just get nutty. But, you know, in the first half of this year, there were something like $160 billion going into venture capital, which is up 75% from the first half of last year, but $100 billion of that $160 billion was all going into AI. Now, a bunch of this was going into the mega deals being raised by the big model players like OpenAI and xAI and so on and so forth, but investors are starting to pivot their focus into the application layer because we're seeing more and more adoption. And so, sitting on top of these LLMs are applications that can, fulfill jobs to be done within the enterprise. And so, while still small, relative to all the money that's been going into the model layer, the application layer has seen the fastest area of growth in terms of funding. Valuations are, quite frankly, they're very high. I've never seen valuations like this, not even in 2020 or 2021. But it's also reflective of the parabolic levels of growth that we're seeing in these companies. There's a number of high profile US based companies companies that have gone from like 0 to 100 million in ARR in 2-3 years...
Paul McKinlay
Executive Managing Director and Co-Head, CIBC Innovation Banking
Thanks (to Dan Lee). Dan (Debow), to the conversation around adoption, especially internally. And some of that like change management, almost like habit building around AI. You know, your alma mater, Shopify, has been in the news a number of times around just how rapid they've been.
Dan Debow
Build Canada, Serial Founder, Investor, and Former VP, Shopify
Up until about 6 or 7 months ago, I worked at Shopify for a long time. And so was there for part of this journey. Look, I think there's a few components of it. The first is, I think, is executives and leaders have to use these tools. So I think that's kind of got to be a requirement because it's very difficult to have problems come to you what you do as an executive and say, “hey, well, why aren't we thinking about doing, you know, using a tool? Can we write a workflow? Can we connect it to a data set that we have so we can answer this question? Can we refactor the work that's being done?” If you have no context or experience using these tools yourself, I think that's going to be much harder. So, number one thing is people have to use them. So second thing, I think we've talked about it, which is you have to create a safe environment where people can tinker. We need lots of people trying and having their intuition of like, oh, I know how to use this. So I think you move to that step. And then the third which was mentioned as well, is like starting to get serious about significant workflows and somewhat demanding of saying, okay, well, like, why are we not taking, you know, increasing cycle time, increasing process. And frankly, in some cases reallocating headcount, where can we say, “Well, we have a thousand people doing this, can we have 100 do it, and do it faster and better?” And those other, folks can do other things that are more valuable inside the organization. And so the question isn't, “how many thousands of people at, you know, CIBC are using the tools every day, but how many more employees, virtual employees have you created?” Because you imagine a world once you start getting adept at using these and setting up a bunch of agents, like everybody gets to hire 3 or 4 people, five people to work for them and do parts of the job, or invent parts of workflow and process that they can then make themselves more effective and get leverage on. And so when you start to think about working with these tools, everyone gets a bunch of employees to work with them, but they're pretty naive. You have to kind of put in 3 or 4 months, just like you would with a junior analyst or intern in the bank, to kind of train them and teach them how you want things, what the expectation is, who you are, what the context is. Once that mindset is set, it starts to change everything.
Paul McKinlay
Executive Managing Director and Co-Head, CIBC Innovation Banking
Dave, maybe we'll go to you quickly around adoption of AI. And some of it would be, you know, if you take the example of Shopify versus a bank or health care system, a more heavily regulated industry. What are some of the balances there around allowing employees to tinker to use the term to, to build adoption versus like the guardrails of, you know, it's a heavily regulated industry?
David Gillespie
EVP, Infrastructure, Architecture and Modernization, CIBC
We've pushed an AI ChatBot with very distinct capability sets to all employees. All of our developers are now using AI to develop. And with both of these, you know, there's no predefined, “thou shalt, do this.” We're leaving it to teams, business users, champions, and we run a very strong champion network. We run all kinds of things to start getting workgroups enabled. And people can, you know, write their own use cases and publish them to the work groups. So, we're starting to really see this take up, you know, so across the bank, about 10,000 of our employees are using it every week, you know, close to 4,000 a day are now on the platform, you know, using it. So, we see a lot of take up, you know, but on these we're not trying to measure. On the other side, we have very specific use cases where we're driving. So, we are now driving into processes where this is the only way the work will happen. And these we have very measured outcomes. We define the outcomes before we take them on. As you're familiar with the bank, we all make commitments. So as a result, you know, when we've deployed the tools we've deployed, there is a number of things people can do with them, which is fine. When they go outside of an ‘individual only’ use case and want to publish something, it will go through a review. And depending on the nature of the review, you know it'll be very quick or it, you know, if you're bringing in a lot of, a lot of client data into something, it would require more fulsome reviews. So, we certainly have review processes built up. So, regulation is key. And regulation, you know, in large part consistently globally is to protect clients in the banking industry. So, with that in mind, like we intend to protect our clients. And we've built out, a number of AI governance around what we're doing in addition to what we had prior to this. As a result, we were quite happy to sign Canada's voluntary GenAI code of conduct, as it was at the level that we're already operating at. And if you put your clients first, you get to the right place. And if you put your clients first, you get to the right place.
Paul McKinlay
Executive Managing Director and Co-Head, CIBC Innovation Banking
I want to double click a little bit on, you know, we heard this concept a few times, from some of the investor community around experimental recurring revenue and how much of sort of the spend in AI and GenAI at enterprises and businesses is kind of tinkering away. We'll see what happens after renewal cycle versus what's really in production and driving real impact. How, from an investment standpoint, Mark, have you guys thought about sort of validating that?
Mark Shulgan
Co-founder, Intrepid Growth Partners
Yeah, I mean, it really comes down to using traditional software metrics just around recurring revenue and how sticky it is. I think your point on renewal cycles is an important one, because a lot of these companies have grown at a really fast pace, and you might not have a full renewal cycle to be able to determine whether or not the product is really sticking. So, the way that we tend to bypass that is by speaking directly to the customers and getting a sense for how much use there is within the tool. And I think use is another measure that you can use to figure out whether or not it's really sticking. There are a lot of instances and you can still hear, you know, just going through this process that, when we were going out and talking to customers of this company, you know, the camp was evenly divided between people who are saying, “You have to be aware of these tools. They're not accurate. They're they're kind of misleading junior employees about, you know, what might be right or wrong and they're just relying too much on them.” And then the other half is saying, “Like, this is incredible. We're servicing our clients much better. We acknowledge that you have to double check the work, but it gets us to a much, much faster, starting point that's really high quality and enables us to serve more customers.” So, we're still at a phase where I think things are being figured out. But, experimentation, I think is, is a really important part of the long term success of AI.
Dan Lee
Managing Director, Technology and Innovation, Investment Banking, CIBC Capital Markets
To Mark’s point, like some of the traditional software metrics like retention don't apply so well here because the renewal cycles haven't really come up yet for a lot of these companies. And so, where investors are double clicking on are usage, right? Usage data across different cohorts. The ROI equation. So, to Mark’s point about ROI and AI, is it our customers measuring ROI against kind of vanity metrics, soft ROI or are there hard tangible numbers that they're measuring? Obviously, the more hard ROI, the more likely they’re going to see a renewal. And then the level of workflow integration depth, right? There's just like a point solution that some of these kind of using off the side of the desk, or is it fully integrated into the rest of the workflow? And so, these are areas that investors have been, spending their time on.
Paul McKinlay
Executive Managing Director and Co-Head, CIBC Innovation Banking
Dan, you've been working on something called Build Canada. Give us a thirty seconds on what you're working on there.
Dan Debow
Build Canada, Serial Founder, Investor, and Former VP, Shopify
Sure, Build Canada is a AI powered civic action group in Canada trying to get, young Canadians and builders, people who build companies, build things, engaged and involved in politics. And I'd say, one of the keys of making the entire program work was, aggressive adoption and use of AI and empowering hundreds of volunteers to build AI powered tools around transparency, government, policy and things like that. So, it's all we're doing.
Paul McKinlay
Executive Managing Director and Co-Head, CIBC Innovation Banking
Two, sort of, closing questions. Dan, like, how do you think about the pro side, the con side and how that shakes out in terms of, you know, productivity, both for companies who adopt it, but also just the economy at whole and, and society as a whole?
Dan Debow
Build Canada, Serial Founder, Investor, and Former VP, Shopify
It's important to remember that what one company does in Toronto or Ottawa about its path towards safe AI is irrelevant to the fact that we're in a geostrategic war right now between major powers that are headlong building, next and next generation capabilities around AI. And it's almost like an open, Manhattan Project that's happening across the United States, in China, and to a lesser extent, other places. And so there's so much capital, so much energy, so much interest in it, because the potential is so huge that rather than asking pro/con, the question is, “When is it coming? And how do I position myself to take advantage of it?” Is a better question for myself. It's not if it will come. So, then the question is, “How do you prepare yourself?” The only logical choice is to move as quickly as you possibly can to start adopting these tools, getting your people to understand the world and work in it, because that is your only best chance when the tide comes. One last comment you said about the nature of the economy. I don't know what it is, but people have a it's like a ‘conceit of the time’. We sort of look around the world and say, well, it's always been like this. No it hasn't. If you had gone to my dad, who's, you know, 80 some odd years, 85, just turned 85 and said, “There's going to be about a half a million people when he was a kid whose job it is, is to help other people exercise. That's what they're going to do. They're going to help other people exercise.” He’d be like, “That's insane! That's not a job. What are you talking about?” I actually think the core skill is for people to become more entrepreneurial inside their companies, inside their lives, inside of starting new business. The tools and friction for people to become entrepreneurial is going to decline. Because, hey, if what I said was true, the average person with an idea who doesn't need to code but can write code now, who can hire people because they hire AI, that means people have potential to build all sorts of new things. As a society, we have to say that's okay. You can take risks. You should try these things. You have to become a society of builders, a society of people who do things.
Paul McKinlay
Executive Managing Director and Co-Head, CIBC Innovation Banking
Yeah. Excellent. Anything to add there Mark?
Mark Shulgan
Co-founder, Intrepid Growth Partners
I think what Dan is saying makes a tremendous amount of sense. My own kind of mindset in terms of how I think about, you know, what I'm what I'm telling my kids is, you know, you've got to be, intellectually curious. You have to be a critical thinker. And you have to execute, right? And I think the entrepreneurship point is a great one. We're going to go through a period of monumental change. A lot of the economic growth that has been driven over the last few decades has been as a result of population change. And so I think now we're going into this, you know, there's a very, huge impact on the economy by a smaller workforce or a slower growing workforce. And these tools are what we need to enhance productivity and ultimately drive more economic growth. So I think, what we're going to go through a period of adjustment in the long run. People are driven by purpose and you find meaning in work. And I think that we're going to find ways to be able to continue to do that.
Paul McKinlay
Executive Managing Director and Co-Head, CIBC Innovation Banking
Yeah. I agree. Well, I want to thank everybody for, for joining us, the fun conversation. Thank you. Dan and Mark and Dave and Dan for being a part of this. And I feel pretty optimistic coming out of this around AI, so appreciate it
