The AI Application Layer - The Good, The Bad, And The Ugly | Jim Lecinski, Northwestern-Kellogg

0:00The CMO Confidential Podcast is a proud member of the I Hear Everything Podcast Network. Looking to launch or scale your podcast? I Hear Everything delivers podcast production, growth, and monetization solutions that transform your words into profit. Ready to give your brand a voice? Then visit IHearEverything.com. Welcome to CMO Confidential, the podcast that takes you inside the drama, decisions, and choices that go with being the head of marketing. Hosted by five-time CMO Mike Linton. In marketing, everything must work seamlessly. If not, efficiency, speed, and ROI all suffer. That's why Quad is obsessed with making sure your marketing machine runs smoothly with less friction and smarter integration.

0:52Better marketing is built on Quad. See how better gets done at www.quad.com slash build better. Welcome marketers, advertisers, and those who love them to Chief Marketing Officer Confidential. CMO Confidential is a program that takes you inside the drama, the decisions, and the politics that go with being that of marketing at any company in what is one of the most scrutinized jobs in the executive suite. I'm Mike Linton, the former Chief Marketing Officer of Best Buy, eBay, Farmers Insurance, and Ancestry.com. Here today with my guest, Jim Leszczynski. Today's topic, the AI application layer, the good, the bad, and the ugly.

1:40Jim is a professor at the Kellogg School of Business at Northwestern, where he has been named Professor of the Year. Prior to teaching, he worked at Google as Vice President of Customer Solutions for the Americas. He's written a bunch of books, Winning the Zero Moment of Truth, and the AI Marketing Canvas, which has just been updated and released in its second edition. This is Jim's second time on the show. Welcome back. Hi, Mike. Thrilled to be a Boomerang guest. There you go. Hey, now, Jim, just to give everybody, our listeners, a background before we dig into the topic at hand. You teach classes on marketing strategy and AI and marketing.

2:25Tell us about your teaching philosophy and how these two topics intersected for you and for your students. Yeah, we've touched on this a little bit in our past conversations, Mike, but I tend to think about, you know, what students need to take away from mine or any business school course or a business school education is a combination of durable and temporal knowledge, right? There's some constants, like they need to know segmentation, targeting, positioning 50 years ago today and 50 years from now. So I try to bring some of that durable permanence, frameworks, concepts, strategic thinking into all my courses.

3:01And at the same time, yes, and there also are some kind of things of the moment, temporal or maybe we might even say perishable kinds of things. You know, if we were going to do a marketing education in the 1950s, you'd certainly need to know about radio. Yeah, that's sort of the new media then. And so here we find ourself in 2025, as we record this, and AI, right, is, you know, certainly all the rage, but it's having strategic as well as tactical and executional implementations, functional process, people implications, cost implications for marketers. So, you know, to have a really complete marketing training, I believe, you know, durable as well as temporal.

3:39I've got to cover both in my courses and strategy. Yeah, I think that's great. Hey, in an earlier chat, you know, you talked about a Gartner report on CEOs and CFOs. And I know this goes into your teaching. So tell us about what you learned from this report, like the key takeaways between, you know, CEOs, CFOs and things they're thinking about. Yeah, I think it's really, really important for all educators and marketers and students to really have their finger on their pulse of kind of like, what are our bosses thinking? Our bosses said broadly in air quotes, right? But like CEOs, CFOs, chief product officers, head merchants, right?

4:23The people who we interact with. And so Gartner does a tremendous job of this. They do a great conference every year where they're sort of, you know, sort of doing some forecasting along the way. And they do quantitative research every year. And so one of the questions that they ask that I keep a close eye on is, what is your top priority? They ask that, you know, your top priority for the business. They ask that of CEOs, CFOs. And every year, year in and year out, this current survey results, 2025 answers no different. By far, the number one priority for chief executive officers and chief financial officers is growth, right?

5:03As my old boss, Eric Schmidt, at Google used to tell us, revenue growth cures all known business problems. When revenue is growing, things are good. We have offsites. We have promotions. We have fleeces and embroidered mugs. I like how fleeces got in there. You know? Yes, because then you get in Silicon Valley, I still have a million fleeces. You got to have a fleece, right? On the other hand, when revenue is not growing, things get ugly fast. And so this is like the CEO's number one priority. At the same time, lots of marketers that I talk to, lots of marketers that you talk to have this notion like doing more with less, cost cutting, productivity, efficiency, like that has to be the top priority.

5:51Now, don't get me wrong, right? Like all marketers, and you know this better than most, Mike, are expected to run a tight fiscal ship, right? You have to be responsible with the money that your CFO, the company money. But here comes the famous cliche, you can't ever just only cut your way to greater glory and profit growth. At a certain point, you need more shoppers, you need more customers, you need more revenue. And so, you know, that's something that I point to often to my students is to say, like, look, never forget the job number one of marketing is to drive profitable, incremental revenue growth.

6:28Yeah, I think that's a really key point that you just made, which is it's not growth at all costs. It's growth that makes the company money over time. And growth that but for what we're doing would not have happened anyway. Exactly. Exactly. So, and so this is a massive disconnect in your mind, this study. Yeah. Between the CEO and the CMO who works for the CEO and then the peer group, which is, we'll just say led by the CFO in the office where they're all talking to investors and everything and thinking we got to grow the business. And the marketer is sometimes thinking, oh, I got to do great marketing.

7:10Is that is that what you're saying? They're thinking? Yeah, I think that's right. Now, you know, I'll refer to one of your terrific former guests, Dr. Kim Whitler, when she was on and she very eloquently talked about the different types of CMO roles. Right. So if so if we're talking about a CMO and a marketing team by her taxonomy, who really is a mark com group, who's given a budget and expected to make ads and that's full stop what they do. Well, then you're always going to be in sort of that do more with less cost cutting because your work has no sort of straight line or direct link to revenue.

7:43So then you're a cost center and you're always going to be under that productivity cost cutting pressure as your foremost KPI. But the flavor of marketing that we're talking about here today, Mike, that you've practiced and preach on the show is marketing as a growth driver, a value creator. Okay, so in that case, we have to have a growth mindset, first and foremost, not a cost-cutting reductionist mindset, first and foremost. And let me just push on this a little bit, because when we talked earlier, you said a lot of marketers, not just the ones that are what's called agency-type marketers or brand-build, just brand-only focused on the Marcom stuff.

8:28You said they're focused on the wrong thing. uh tell us what you mean by this well so now we talk about you know the famous means and ends conversation right so so there are um business outcome metrics revenue profit share customer count footfall right retention yeah yeah those kinds of retention average order value average basket size and then there are means uh metrics we'll call these activity metrics so this is time spent on site per session, bounce rate, click-through rate, cost per click, right? So so much of like when I ask marketers to show me their dashboard, it's these sort of intermediate means or activity metrics.

9:14Right now, I mean, I guess all things being equal, I probably would like to pay less rather than more for a click, but you don't take click-through rate or CPCs to the bank. And you don't talk about that in a shareholder meeting or with your board, right? Those are means to driving revenue, profit, share, right? The business outcomes. And so I think it's just important, you know, to not forget tail and dog or dog and tail and confuse the two. I see what you mean, but, but I do like the, I do think there's a lot of false precision in that, which is we can't measure maybe everything, but we can measure this, and we will just tear it apart and do false precision and make it better every time.

10:01Wow, we might be missing the big story. There's also other things that marketers are doing on culture front, everything else where they're trying to tell us about that too. We are taking a quick break from our discussion for a word from our sponsor, Quad.

10:44and smarter integration. Better marketing is built on Quad. Now back to our discussion with Northwestern Kellogg marketing professor, Jim Lesinski. Yeah, so, I mean, again, we say if marketing's role, first and foremost, is drive profitable incremental growth. You know, at Khan, we've got lots of marketers kind of sitting there saying our job is to be part of the culture, right? To be water cooler talk, top of mind, these kinds of things. And I would take some friendly or not so friendly issue with that mindset. Again, means and ends. If being more top of mind, we all measure aided and unaided, brand awareness, attribute awareness.

11:29So sure, I'd rather be more top of mind than less top of mind. But I don't take cultural insertions to the bank. I take revenue to the bank. So, you know, when I hear marketers say either our job is to cut costs or our job is to be popular, you know, maybe. But, you know, on all of these things, you know, during my time at Google, we would regularly see marketers who actually counterintuitively by increasing their cost per click, we're actually able to find more valuable customers with a higher customer lifetime value and actually drive higher profitability. So, you know, got to be careful not just get caught up in, as you say, the things that we measure because we can.

12:11Yeah, the easy things to measure and the things that you can control at this level. And the other thing that always has bugged me about a lot of those dashboards you talk about is all those things really stack on each other to drive sales and profit. By maximizing one, it doesn't necessarily mean you're going to do anything to the others. Yeah, that's right. If you run a lot of traffic, you can buy traffic. If it doesn't buy anything, it's a worthless metric. That's right. Now, there are some linkages that we well know, right? Like, you know, increasing, well, as it's famously called, mental and physical availability, but like increasing distribution, ACV, PCV, right?

12:50Points of presence. Like, we know that that typically has a pretty straight line. And ACV is all commodities volume, which would be shelf space. I like how you just threw in all those consumer goods. acronyms you just popped them out there good so how many how many marketers percentage-wise in your mind whole all the marketers in the country or you can go to the world we have like we have actually a pretty big international audience are missing the mark on this right so if we set aside you know the portion of marketers who are doing sort of the marcom marketing equals advertising tasks so now we're looking at you know by kim whitler's measure i don't know half two-thirds of all marketers that are in the revenue growth mindset, kind of a function of marketing.

13:37You know, I would say, again, not a poll, not scientific research, not academic research that I've done, but, you know, in my consulting, in my conversations, in listening to these marketers talk, I would say probably we've got maybe a third of marketers who get caught up in these kind of activity, intermediate metrics, cultural kinds of things, not the revenue kinds of things. So, you know, I'm curious, Mike, what your take would be. I think a third is even generous. I think there's actually more. And I also think you would not see this in the CFO thing where people would say, I don't care about forecasting or I don't care about reporting.

14:19I only care about this part of it. So I think that is a good call out to marketers. um let's talk about the topic dominating business for the last couple years ai you think there's a good story and a bad story here uh the good is that marketers and companies are embracing ai and the bad is that they're focused on just the efficiency and time savings and cost cutting tell us what you mean by by that whole little run of words i just put out there yeah so this sort of just flows from a previous part of our conversation where we were talking about how How many marketers are focused on, you know, doing more with less cost cutting, productivity, efficiency and saving, like minimizing yourself versus the means versus the transformational value creation, the growth and the ends.

15:11So now is it a surprise that when sort of this amazing new technology comes along, marketers would carry that mindset that they had with them? Am I a productivity minimizer or am I a value creator maximizer? they're going to carry that mindset with them when it comes to AI. So business school professor, so we think in two by two. So a way to think about this then is like, what's the use case for AI? Well, we could think about the benefit on one axis and the beneficiary on the other axis. So the benefit could either be, hey, why are you using AI? I'm using it to be more productive and more efficient, as we've been discussing, or I'm using it for value creation, for transformational growth.

15:55That's sort of the benefit on the horizontal axis. And then who's the beneficiary? Like who is seeing the benefit of this? It's either internal, something just for my marketing team behind my firewall, but in my physical office space, that's never seen externally by my customers. Or is it internal, sorry, or is it external, something that my customers see? So now Now, if you sort of look at that two by two in the bottom left-hand corner, we have internal productivity. This is using AI to summarize Zoom calls. This is using AI, right? Yeah. Did that before. It's the same output. It's just - Automatically doing the charts for me, automatically doing this.

16:40Yeah. Right. Write the first draft of my creative brief, still the same creative brief. It's just done in six minutes instead of six months, right? Right. So so many marketing teams that I talk with start and are anchored. And many of them, frankly, sort of see that as the limit. The end game of A.I. is just this internal productivity. Same stuff we did before, same way we did it before is just cheaper. I mean, this is Henry Ford's faster horse. Yes. On the other hand, if you go to the upper right hand corner of that two by two, this is sort of the optimal axis. This is value creation, transformational growth by using AI to enable and empower your customers, your users, your consumers to do amazing things.

17:29Example, weirdly, a lot of the apparel companies are doing a really good job here that you normally might not expect. So Stitch Fix, Ralph Lauren doing some amazing things here. What are they doing? Tell us what they're doing. Yeah, so Stitch Fix is doing some cool things here. So they have obviously e-commerce company. They have a lot of data. So they know every single page I've ever looked at, every single sweater or pants or top that I've ever browsed. They know everything I've bought and they know everything that everybody like me has ever looked at. And you fill out a style quiz as well. So I give it some like direct training.

18:07We call that training data right along the way. So now what they're able to do with that is to create a virtual or synthetic personal style consultant, much like, for example, the human, if I walked into Nordstrom, right, who would walk from section to section and help me find what I need for a job interview to wear or a wedding. But now I'm able to then interact with that. It's an app in Stitch Fix and Ralph Lauren's case. That one's called Ask Ralph. Your listeners should all just try this to sort of see what this looks like. But now I can go in there and say, I just got this wedding invitation for a wedding in Hawaii.

18:45And it says casual to semi-formal beach attire. Oh, I love that. Or casual business resort wear. Right. I have no darn idea what that is, but I'm pretty sure I don't own that in my closets. So, you know, now what do we do? I can ask it to help me out. It knows all of this zero and first party data information about me and people like me. It can go and assemble a personalized, customized, real-time recommendation on the fly that's adding value to me as a shopper, increasing CSAT, increasing customer lifetime value. I mean, you can see, Mike, how this is like more money for the brand. Oh, yeah. And I'm just – I have a vision of you in customized Bermuda shorts right now.

19:34So if you want to send us a picture of what you're wearing to the beach resort casual wedding, we will absolutely post it. Yeah. But you see, like, so this is sort of the two poles. So if I'm a marketer who's in that cost-cutting, productivity reductionist mindset pre-AI, and I bring that mindset with me, well, then sort of the vision, outer limit of my vision for AI is just like, you know, maybe we could use it to save some time and fire a few people. versus, right? Oh man, look, this can actually help me grow the business not 5 % next year, but 6%, 7%, or 8 % next year. Right, and then the first people doing this get all the experience too.

20:22100%. So behind this statement is that there's a lack of vision from a lot of folks out there or that people just aren't thinking about this right. and how do you how do you fix that like is that a person thing or is that a culture thing or what is that like what what do you say you need more vision or you're thinking about it wrong or um how do you how do you fix that kind of problem yeah i mean kind of all of the above right so i mean what did malcolm gladwell teach us whenever any disaster happens it's not one thing it's a thousand things yeah right so we have to sort of unpack for each brand and i you know i do this kind of consulting work outside my teaching.

21:05I work with brands to help them, you know, transform, we call it from zero to superhero with AI in their marketing department. And so, you know, first is we have to do some actual computer science training and education. Most marketers didn't come with a comp sci background. And so to really help them understand how these tools work, what they're capable of. And again, that doesn't necessarily mean, although sometimes we write Python code, but it always doesn't mean that. It's just helping to understand that these are prediction machines. They can either predict or they can generate. They can handle probabilistic questions very well, but deterministic questions where there's only one right answer, not so well.

21:46So if I ask it, what's the population of Portugal? It's not going to do as good a job as I say, what might be three good new business ideas? Those are probabilistic questions. So we've got to help them understand like the technology of how this works. Then from there, we help them understand the four use cases in those two by two that we just talked through, show lots of examples of lots of brands, good, bad, and ugly, right, that are doing these kinds of things to expand their understanding and imagination of what's possible. And then, you know, from there, and we can talk about this a little bit, you know, we have to help them understand some of the risks and avoid the bear traps along the way.

22:24And then the last piece is, okay, how do I get started? What's a do on Monday, do next month, do the month after, do next year, kind of a maturity ladder. You said you think there's five risks that are in the way of the vision thing. Yeah. Privacy, accuracy, regulatory personnel and reputational risk. Tell us about these and how marketers should manage them, because it's almost always easy to see the risk. Managing it is another thing. Yeah. And I think it's it's I'll unpack these, but I think it's because of these risks is why a lot of marketing teams limit themselves to these internal productivity things.

23:04Right. Because, well, if I try something and it's not quite right for my team, I'm not going to get fired over that. Right. Right. On the other hand, if I put a chatbot out there and it hallucinates and goes crazy. Right. Now I got like my board and my CEO calling me at home on a Sunday. Yeah. So. So, yeah, I mean, privacy risk by data privacy. So if you think about I've used this term before zero party and first party data, right, either things that your customers volunteer to you or you are allowed to gather throughout the course of a transaction. I need to bring that context to any of these large language or multimodal models.

23:38They know everything that's out there in the world on the Web, on Reddit, in the New England Journal of Medicine, but they don't like know my customers and my business. So I have to give them that context. The challenge or the problem here, though, is unless I'm using sort of an enterprise class, enterprise grade model like ChatGPT Enterprise for big business, ChatGPT Teams for medium sized businesses. right if i've got my employees just going home at night using what's come to be called gray market ai consumer chat gpt because my team and my cio has not given them state-of-the-art tools that are safe to use guess what happens they go home they've lob all this stuff into regular chat gpt and then the data leaks right at worst at you know the other problem is it trains the underlying model on my IP.

24:30So this is this weird thing where lots of CIOs are rightly moving cautiously, right? They say, we have to be very careful with our data. We have to vet all these tools. But the longer the CIOs wait to empower their teams with state-of-the-art tools, the more they're using gray market AI and actually exposing the risks that the CIOs were trying to prevent. And falling behind in the market. Not to mention falling behind. So that's the first one. The second one is sort of a legal or regulatory risk. So lots of questions about, you know, the intellectual property of what was used to train these large language models.

25:09In fact, just in the news this morning, I don't know if you saw this one, Mike, but, you know, Sora 2 is now out and the Motion Picture Association has sued them because now you can ask it to create a Mandalorian, right? Which is Disney IP. So, you know, if I do that as a brand, if I use these tools to create something, am I actually at risk of getting sued because I use the output? And then part B of that is, can I actually copyright it? Because we always copyright 2025 Best Buy Incorporated. Okay. Well, in the US, and I know we've got a lot of global listeners, so different rules in different countries, but in the US, the US Patent and Trademark Office and courts have ruled that unless something is substantially created by AI, substantially created by humans, you can't copyright it.

26:00Right. Now, we don't know. Does that mean 90 percent, 51 percent? That hasn't been tested yet. Yes. But substantial feels like it has to be 50. Yeah, I would agree with that. But we don't like we don't know yet, although I would say I've had this conversation with some Fortune 500 chief legal officers in a friendly way. And I say, Mike, if there's just a little lifestyle image that's going to be used one time once in an email, do we really care if that's copyrightable? Yeah. You know, my pack shots, my product, my store photos. Sure. But I don't know, like if this is just an ambient side thing in an email once, maybe we don't care.

26:39But that's the second risk. Third risk, of course, is now personnel risk. This is the who moved my cheese. You know, Jim Farley, CEO at Ford, has come out and said the quiet part out loud, that he expects half of all white collar jobs will be severely disrupted or eliminated in the next decade. I love that word severely disrupted. Well, I mean, we've seen this in the world with blue collar jobs over the last 25 years. Again, it's technology. It's not AI that's disrupted those blue collar jobs. It's robotics and automation on assembly lines and factories. But that's coming, right? So that's coming here.

27:14So you need to know as a CMO, if you stand up and say, hey, guys, we're doing a big AI project, half the employees will say, oh, you're just trying to fire me. Yes. So, you know, that's an issue. You know, reputational risk is now hallucinations. Air Canada very famously put a chatbot out there. A poor soul said, right. Hey, do you have a grievance fare because I want to fly at a discount for to a funeral? It hallucinated because that's a probabilistic problem. Most airlines probably have such a fare. So it assumed so. When he went to book the ticket, the human agent said, sorry, we don't have this, went to court, the court said, guess what?

27:52You own it. You own it. So that's a hit to their reputation. And then the last one, of course, that we talk about here is sort of deep fakes. And now, you know, so many brands are using influencers and micro influencers. Well, the micro influencers are now famously using AI to generate images of themselves in the VIP seats at Wimbledon, and they were never even there. Okay, so you're paying them to be an influencer, and now they're sort of fake influencing, right, and not disclosing that. So what does that mean for you as the brand team then, right, because they'll flame back on you. So there's mitigation paths for all of these.