Branching Out by Bonsai | Matt Seitz on AI, Education, and the Future of Work

‍

Hi, welcome back to Branching Out. This intro was filmed on another day, but I'm really excited for you to talk to Matt Seitz. He's the AI Hub Director at the University of Wisconsin's Business School. He's also my former boss and colleague when we were both at Google. I'm really excited to hear his perspective on the current state of AI, the future state of AI, and how you should all use it in your work and in your day-to-day lives.

A couple of other announcements that are coming. We have a couple of big things in July I'm really excited for you all to see. We will have additional Branching Out episodes. We will have a lot more video content for you coming around our product, a bunch of new releases that I'm really excited for you to see. And hopefully some fun stuff along the way. So check out this episode. Matt's a great guy and a good friend. Hope you all enjoy it. Cheers.

Matt: Thanks for doing this.

Matt Seitz: Yeah, it's a pleasure.

Matt: So let's see. How long has it been since you took the new role at Wisconsin?

Seitz: It's been 5 months.

Matt: Okay. So, it's new.

Seitz: So, it feels like yesterday to me, but it's super exciting and appreciate you coming down. Love the Badger gear and Big 10 all the way.

Matt: When I talk obviously about Bonsai, people are very often looking to hear what I have to say about Google. I do want to talk a little bit about our shared time together at Google, but just to indulge me, I would love to hear a little bit about your career before Google. I think it's really interesting. Can you tell us a little bit about what you were up to before we worked together?

Seitz: Well, first of all, it's a pleasure to be here and spend time with you again, Matt. To spend time here at Bonsai and to see what you're building is really neat and distinctive in the market.

I think I'm at year 30 of my career. Pre-Google I was a techie. I was a programmer out of college, and then I did IT project management, and then I got into data after that. Consulting, McDonald's, Abbott Laboratories—so a mix of industry and consulting. I think of myself as someone who loves technology—I'm a nerd—but I also like working with people. That's the thread that I think wove through all of it and eventually made it to Google.

Matt: The people who know you really well know about your programming background. But if folks are meeting you for the first time, I think that's something really interesting. How do you see having that kind of foundation influence how you thought about your future consulting career?

Seitz: The way I think about it is I love technology. Technology is at the core of disruption in business. For me, it was the PC, then the internet, and then the smartphone, and now AI. I think AI will be the biggest disruption of any of those four for our world. In another way, it builds on all those four. It's a special moment. I think the chance to be a part of that is really special.

Matt: It's so cool. We clearly worked at Google together. Someone's always going to ask you what made you pivot from a hugely successful time leading analytics at Google to now you're in academics. Tell me about that.

Seitz: There are a few things. Broadly, I mentioned I'm in the 30th year of my career. So the chance to be more mission-oriented is really neat. What do I want to do for the next 10 to 15 years? I want to make sure I'm spending that time in a way that is driving things I really value.

There are a couple of elements that I think are neat personally. I grew up in Champaign, Illinois. My parents are professors at Illinois. I went to high school on campus. So the chance to be in academia again is really rewarding in that way. And then in terms of Wisconsin, my daughter goes to UW, so it's fun to share that with her.

Matt: That's the ultimate dad hack, by the way. Did you somehow find a way to say, "Hey, hey, how you doing, honey?"

Seitz: We have this running joke. In her apartment, one of her roommates moved out in the fall. And so, I said, "Hey, Danielle, this is great. I can just stay with you when I'm up here." She wasn't on board with that.

Matt: Did I get your text messages read after that one?

Seitz: "No, Dad. I'm out of town." She wasn't buying it.

Matt: We were together at Google in like 2010 through about 2020ish. You led a lot of teams, including the Retail Analytics team I was on. Is there any insight that you brought to our senior leader customers that stands out to you as really surprising or something that really stuck with them?

Seitz: It's remarkable working at Google the power we have at our fingertips as an analyst, which is primarily Google Search—how many people are searching for X at any time. What we did—and you were a huge part of this, Matt—is we stitched that demand together with the fact that because of Google Shopping, Google knows every product that's sold virtually in America, along with who's selling it, do they have it in stock, and what the price is.

We'd sit down with a CMO or CEO and say, "Hey, by the way, you want to be a leading seller of toys, but you're not selling this Barbie Playhouse that tons of people are searching for. Do you want to sell this product or not?" That was very eye-opening. And then the other thing we would add on is location data. So you could say, "Here's your store footprint, but based on what people are searching for, you should be in these three markets or maybe you're over-stored in these markets." It brought a texture to the demand data that was super relevant for retailers.

Matt: I remember a lot about working on those projects with you and agreed. It was one of the most fascinating things. You got to give Google a ton of credit. There's an attitude at Google that data is an asset, not something to be scared of. That's a culture thing that is really powerful.

Was there any big misconception that you felt was consistently presented to you from the clientele side that you maybe still see today?

Seitz: In the marketing space, what I found was that marketers are so used to having the hood up and tweaking keywords and bids. "That's my job—to tune this engine of Google to get the best performance." But AI is better at that than you, a human, can ever be.

Having a marketing department shift from saying "instead of doing that, I'm actually going to understand the business" is the shift. I'm going to understand what kind of customers add the most value, which SKUs if I don't sell more of I'm going to be scrapping at the end of the season, and using that to direct the AI. That is a shift that I think marketers are still struggling with, but is where the marketing department of the future will really add value.

Matt: Hypothetically, if you had stayed at Google, what would you have worked on?

Seitz: Probably helping customers deal with the shift. There's a lot of anxiety. The search results page is being overtaken by AI overviews. Google just launched AI mode, which means maybe people aren't even going to search—they're going to use a chatbot interface. How do I think about Google as a driver for my business? That would be the primary conversation.

A dream job is in AI. One of the races I'm excited about is the race for the ultimate personal assistant. I think Google and Apple are the two that have the most customer data that could do this. Google has Gmail, calendar, maps, photos—incredible assets. And then, of course, Apple has all your on-device stuff. I would love to be in the middle of that. Finding a way to bring that to life because it's so powerful.

Matt: I don't know if I would characterize it as a misconception, but we were at a meeting with AI tinkerers in Chicago recently. Someone asked about the AI experience in Google Search and shoppable experiences. Will marketers need to advertise anymore? It was a great question.

AI is incredibly impactful and will add value, but consumers want to shop. That's why they're shopping. There's going to be an exchange. AI is going to play a role in making that better, faster, but if people didn't want to shop, the Amazon homepage would just mail us stuff. We don't really understand how to use it yet or why it's valuable.

Seitz: What's remarkable is if you look at the space right now, you have OpenAI, Anthropic, Google—everyone's trying to be the leader in the technical space. But then you have all the businesses—Walmart, Target, AT&T—trying to figure out how to adapt. Then you have the startups. At the same time, the fundamental technology is still changing. Google recently released V3, which is a material advancement in video generation. Everybody's trying to operate their business while the underlying technology hasn't even settled yet. It's a remarkable time.

Matt: You're now shaping AI curriculum at Wisconsin. Tell us a little bit about what your new role is and what you think AI fluency looks like for non-technical leaders today.

Seitz: What's neat and a challenge at an academic university system is AI touches almost every portion of the system: what you learn, what you should learn, how you do research, the pace at which things are changing. At a business school, AI touches every single business function. We have to operate and adapt in almost every part of what we do.

We're creating this thing called an AI Hub. How do we put people together? Put industry together with students, researchers with faculty. How do we make those marriages to create value in meeting the moment as an institution? It's a chance to build.

Literacy and fluency is remarkable. ChatGPT launched less than three years ago. If my daughter is a junior at Wisconsin, it launched after she started college. It's still so new that I'd say there's a whole bunch of people that know a little bit but not much.

Matt: Your daughter's getting a ton of play on this podcast. She can tell her friends, "I was in college before I could ask ChatGPT to give me the test answers."

Seitz: That's right. It breaks some of the fundamental constructs of learning. When you and I went to college, we had to write an essay or do calculus. It was not fun to write an essay—it was hard—but that's how our brains learned: through that struggle. Now you can just type a question and get an essay. It breaks how you learn, and on the other end, it breaks how you assess. Did this student learn what we were supposed to learn or not? We have to have a new model for learning and deciding what they need to learn.

Matt: You're a triathlete and a runner. There's a ton of training involved. You mentioned the struggle. Do you relate any of those experiences to your job as a teacher and your students?

Seitz: Don't sell yourself short. For those watching, Matt was a Big 10 baseball player at Michigan. He's an athlete in his own right. A legend.

Iron Man is hard. Really hard. Some things I've learned through the process: One is setting an ambitious goal gives you a focus. I had never done a triathlon before and COVID hit and I'd always wanted to do Iron Man. I signed up, paid the $1,500. I showed up to train with this group and they're looking at me like, "What the hell?" I had my helmet on backwards for the team photo. But every day it was, "What do I do today? What do I need to learn?"

Setting a big goal focuses you and stretches you. The other one is doing it with others. There are a lot of long days in training. Doing a six or seven-hour bike ride alone is much less engaging than going on a ride with somebody else. Finding people that you can work together with makes it a much more rewarding experience.

Matt: The difficulty of the objective is what makes it fun to go after it. We all have fears and that can take time to teach people. I hope that AI doesn't rob people of the experience of struggle. I do sense that there are new struggles coming though.

With AI doing all this stuff for us, tell us your opinion. Where does humanity outpace AI by a lot? We know that humans shouldn't be adjusting keyword bids anymore, but what should we be doing?

Seitz: I get this a lot from students: "What should I study? What should I focus on?" One answer is: we don't know what the end of the model is going to be. They're still improving very rapidly. Some people think all teachers and doctors will be out of work. I'm not there.

Think about somebody who's building a marketing campaign. You can say, "Hey chatbot, go write me a marketing brief." It can write you a brief, but what it can't do is understand what your agency thinks about it. It can't understand what finance's goals are. It can't understand what your merchants care about and what happened last year and all the context you can bring into that conversation.

The marginal value of producing content may go down, but all the things around it are going to get more important. I talk about synthesis. How do I take something here that a chatbot can create or an image here, work on top of all that, and put those together into something that's now better than just something I could ask for? It's a little bit of cognitive offloading.

Matt: I appreciate the non-magic of technology. It makes it feel like something that we should be able to do.

Seitz: Two things to think about there. You should understand what these chatbots do. They seem magical and people want to say they are sentient beings. But really, they are an average of all the text that's ever been written on the internet. What they're doing is saying, "Based on the text you just wrote me, what is the most likely response to that based on everything I've ever read?" When you start to understand it's not the Terminator, it's really helpful.

The other thing is a key skill we will need to have is the ability to evolve with new technology. Mary Meeker dropped her AI report earlier this week. One of the stats that's in there was it took the PC 20 years to be adopted (50% adoption in the US). Internet was 12 years. Smartphones, six years. And chatbots, three years.

The pace of change is going to accelerate. That ability to incorporate change behavior is going to be a key asset for anyone in the economy going forward.

Matt: So we're getting to the fourth quarter of this interview. Are you ready for "Jump Around"?

Seitz: I'm ready for "Jump Around."

Matt: A few fun questions just to wrap this up. What is a book every future leader in the space should read?

Seitz: I love reading presidential biographies. I'm going one by one starting with Washington. I think I'm going to do Taft next. FDR, Lincoln obviously. There's a four-volume biography of LBJ that is incredible. If you're not into politics, read Catherine the Great or find someone in the domain you care about. Biographies are excellent learning opportunities.

Matt: Favorite AI tool that you personally use?

Seitz: I like the chatbots, they are amazing. But I am able to code again thanks to AI programming tools. Cursor is just magic. It's like this autocomplete and it goes and fixes code five lines down that it knows it should fix for you.

Matt: Is there a very underrated AI tool in business today that you feel like people don't talk enough about?

Seitz: A skill people don't use is personas in chatbots. You're going to go meet with an executive. Tell your chatbot: "Hey Claude, hey Gemini, you are X. Read this email. Read this presentation and give me feedback on it." It's really valuable.

Matt: Final question. Coffee, cardio, or code? What are we doing?

Seitz: Cardio. Yeah. If you want to do Iron Man, you have to train before the start of the day. You can't wait to try to do it at lunch. You got to be at 5:00 a.m. getting it done. There's something good about starting your day and being like, "Okay, I just did an hour and a half ride," and then you just feel invigorated for the rest of the day.

Matt: Before you go, you run a pretty cool newsletter. Can you make sure to plug it?

Seitz: It's called Explain It Like I'm Busy. The goal is to consolidate everything I spend 15 to 20 hours a week keeping up on into something you can read in three minutes. And I try to make it fun. I put images and I have a dad joke in there. Find me on LinkedIn and you can follow it.

Matt: Thank you for coming, man. It was always a pleasure working with you and getting to learn from you.

Seitz: Thank you, Matt. It's a real pleasure.

‍