← Insights Unlocked
Insights Unlocked
Leading AI transformation with Greg Shove of Section
00:30:06
Greg Shove shares how leaders can drive AI transformation, overcome adoption barriers, and unlock AI’s potential in business and customer experience.
Loading summary
Transcript37 lines
- [00:02]
Nathan Isaacs
Welcome back to the Insights Unlocked podcast. In this episode, we're joined by Greg Shove, CEO of Section, who shares how a single hour of ChatGPT changed the course of his company and his thinking on leadership in the age of AI. From boardroom simulations to building an AI first culture, Greg offers practical insights on navigating fear, driving adoption and leading transformation. Enjoy the show.
- [00:27]
Michael Dominick
Welcome to Insights Unlocked, an original podcast from UserTech where we bring you candid conversations and stories with the thinkers, viewers and builders behind some of the most successful digital products and experiences in the world, from concept to execution.
- [00:46]
Nathan Isaacs
Welcome to the Insights Unlocked podcast. I'm Nathan Isaacs, Senior manager for content production and user testing. And joining us today as host is Michael Dominick, User Testing's head of AI. Welcome, Michael.
- [00:58]
Greg Shove
Hi everybody.
- [00:59]
Nathan Isaacs
And our guest today is Greg Shove. Greg is the CEO of Section, an online business school for the age of AI. He is a six time founder, investor and AI thought leader. He helps executives leverage AI as a strategic partner, driving innovation and decision making. Welcome to the show, Greg.
- [01:18]
Greg Shove
Thanks guys. Great to be here.
- [01:20]
Fantastic, Greg, it's great to have you. So look, you've been a founder multiple times and have seen major tech shifts firsthand. So what was the moment that made you realize generative AI wasn't just another trend, but something that could fundamentally change how businesses operate?
- [01:39]
Yeah. Well, Michael, first of all, let's go back to February 1, 2023. That was the moment I was playing with chat, GTP, chat, GPT plus and I was like, wow, this could be what keeps me working. I want to keep working. I live and work in Silicon Valley, the most agiist work environment in the world. If you're not 28 in a hoodie, you're an idiot. I'm 63, I type with two fingers. And in fact, I've started seven companies because I started one more last year. And I just thought, okay, this was going to be my cognitive edge, the way that I would sort of keep me sharp and keep working. And it was kind of that hour actually that day, that hour that I decided to pivot my company, Section, which I'm the CEO of, to become an AI school for the age of AI. And so kind of all in from that moment on.
- [02:25]
So, Greg, what. What were you doing? What did you say? February 1st of 2023, like what, what were you actually? What was your prompt and what was the output?
- [02:34]
Yeah, great question. I'm not sure I remember probably a bit of both like work and personal, so I can imagine. I was impressed in the early days with kind of thought partnership with AI asking, you know, medical questions or travel questions, things like that. So, you know, life at home. Right. And then life at work. One of the early use cases that we used at work was we asked AI to pretend to be a board member and we sent our board deck, the deck that we send to our human board members we would also provide that we still do. We've done this every quarter actually since then. We provide the board deck to ChatGPT, Claude, Copilot, Gemini, and basically ask the AI to adopt the Persona of an aggressive board member, of a conservative board member, you know, and really kind of get us ready for board meetings. And as I said, we've done it every board meeting since the AI's score, about 85 to 90% of the humans, meaning we compare what the humans tell us in the actual board meeting to what the AIs told us before the board meeting. And you know, AI is getting most of what the, what the humans are telling us. You know, not everything. And by the way, AIs don't write checks yet. You know, you still need board members, at least if you want the money, but if you're a startup. But that to me was a pretty eye opening use case, particularly for leaders. I think a lot of leaders think that they don't use AI. AI is for the content marketers or AI is for the software engineers that other people do. That kind of hands on work that needs AI. But that's not my point of view. My point of view is any medium to high stakes decision. Every executive, every leader in any industry should be using AI at work as a thought partner and frankly at home as well in terms of medium to high stakes personal decisions.
- [04:20]
So I've heard you tell that story before about, you know, using AI in a board meeting. And I've also heard you talk a lot about, you know, how executives should be using AI a lot more, but maybe they're not. Not, not as much as we would like to see. So what are the barriers? What's stopping those individuals from fully embracing and integrating this?
- [04:41]
Yeah, I think, I think a lot of things, Michael. I think that maybe not executives, but, but maybe, but most of us have a fair degree of anxiety and sort of concern about AI, understandably. Right. And all the research, and there's just, there's some new research just this past week from Pew Research about the levels of anxiety about job loss, about, you know, job degradation and so on. And so again, not surprising if you get your news about AI from the Media, the media's job is to outrage us and scare us and get us to click. And so really the narrative about AI has been this AGI, this superpower, you know, this sort of, this coming agent that's going to take our job. And by the way, you and I were both at the, the AI conference last week in Las Vegas, Humanex, you might have noticed that one AI vendor on the, on the conference floor, you know, stop hiring humans was their sort of tagline, right, right across the front of their booth, you know, and Mark Benioff at Salesforce is talking about agent force all the time as a replacement, you know, to humans. So not surprising. People, people are anxious. That's one. They're not getting properly trained. It's clear. And again, we've looked at research and we've done our own research at section about employee satisfaction with training, about AI, and there is no satisfaction about the training, meaning most employees aren't getting trained properly. I think most companies, Michael, deploy this thinking it's software. I think it gets bought by the head of IT or the CTO or the head of technology. And I think there's this misconception that AI is like software, like an ERP or a CRM or some sort of marketing automation software. And it's nothing like software. It doesn't behave like software. We don't interact with it like software. You know, it hallucinates, it gives us different answers to the same question and so on. So I think that just slows adoption. It is kind of, there's not a real appreciation for what we're provisioning to employees is, is some sort of a teammate or co worker or, or thought partner. And it won't always generate great results. Some days it does and, you know, some days it doesn't. I'd say the final thing too, Michael, is overly restrictive AI policies and guidelines, particularly around data. AI to me is not really useful unless you can use your data, whether that be at home in terms of your own medical information, your own medical records, or at the office. You want to load in your business plan, you want to load in your marketing plan, your product roadmap, and then work with AI, have a conversation with AI to help you answer the questions or work on the tasks that you need to accomplish. And so I think a lot of companies have had this somewhat irrational, frankly, fear about data privacy and kind of corporate data. And so our employees are scared to use AI and certainly scared to load up their company information, which I think robs AI a lot of its potential. And Sort of power.
- [07:32]
Yeah, I agree with you on all of that. And at Section, so you're offering businesses content courses that are going to help their employees upskill and move through this transformation. And I should mention that User Testing and full disclosure is a customer of Section. We are using your content for the same purpose. Right. To help us enable that transformation. So kind of putting the focus on your courses. What are those in demand AI related courses that your students are enrolling in? What do you see resonating with businesses with the content that you're offering them?
- [08:08]
Yeah, I think it's probably what you would expect, Michael. And I'm sure what, what most of the employees at User Testing will take advantage of as well, which is if you're new to AI, you want to get the basics so you know, basic prompting and how to get the most out of your conversations with AI. And then of course the advanced version of that course. So basic and advanced pretty quickly after that. Employees want to make it relevant to themselves. So they want to take the course about how to build your own use cases, how to kind of do your own workflow audit, if you will, on yourself or on your team, and then find the use cases where there's the highest return to be using AI. And that means basically functional courses. So AI for marketers, AI for finance people, AI for HR and so on. So we have a growing catalog of functional classes as well and then finally some of those leadership and strategy classes. So if you're leading a team that's implementing AI, how do you do that and how do you change the workflows of the team? How do you get the team ready to adopt and deploy AI? And then, then how do you set AI strategy? So we have one class that's really sort of a classic strategy class for leaders. So, so they can figure out a framework to prioritize and greenlight AI initiatives.
- [09:20]
Yeah, so you have a pretty good look at what businesses are doing and where this is being adopted within businesses. So is there a specific role type that you're finding is leaning in more than others? Is there a specific part of businesses like marketing or sales or you know, engineers that are leaning in more towards this transformation and the content through Section?
- [09:42]
Yeah, yes, yes and yes. I mean it's all the language intensive functions. Right. So we, we know it started with software engineers and I consider them, Michael, the canaries in the coal mine. I think as leaders we should all be studying carefully how software engineering teams are using AI and they're kind of the coding co pilots, whether that be Claude, you know, sonnet 3.7 or a copilot like Devin or Klein, these kinds of AI tools for engineers. Because what's happening there is going to happen in my opinion, in the rest of knowledge work in the coming couple years. And so certainly it's language intensive functions. So that means software engineering. You know, code is a language and then it is marketing, it is sales, it is obviously customer success or customer service. You know, the idea of chatbots replacing humans and, or augmenting humans for customer service is an obvious use case that, you know, that's a language intensive function. Quite frankly, it's function where we trained humans to be more like bots, which is customer service agents. Right. And manage to kind of, you know, quantity targets and, and time to resolution targets. You know, really we kind of mechanize that work. If you think about it, it makes sense to me that that's an obvious place where AI is going to augment and then really replace, you know, human work. So it, it is those language intensive functions. Typically we see HR finance and the leaders lagging in terms of AI adoption. That makes sense. They're more risk averse and again they're not as language intensive. But I would say with the recent releases of products like OpenAI's Deep Research and the more broadly the reasoning models that these frontier AI labs are releasing, I think we'll see that this AI is getting good at math and other functions like the HR and finance functions will, you know, we'll get real gains from AI.
- [11:34]
Yeah. So one point that you touched on a little earlier that I want to put a little bit more of a spotlight on is the fear and hesitation that folks have. So businesses like user testing and businesses like those that you work with at Section, we're pretty serious about AI transformation. Right. But we have a lot of folks in our organization who are understandably and justifiably, you know, a bit nervous about all of this. So what is the right response for organizations and leaders who are serious about AI transformation? Like what is the right balance between pushing for that transformation while recognizing that this could be disruptive to roles?
- [12:12]
Yeah, I'd first of all say, you know, every company and situation is, is context specific. If you're working in an industry where AI is not yet going to be widely deployed, maybe because of regulatory reasons, you know, you've got more time if you work in a language intensive function or business in a language intensive industry like software development, you know, AI is coming, you know, faster than you think. So I do think part of this is being intentional around the pace that you want to go as a leader with your organization. You know, I'm okay if people say intentionally I'm going slow, you know, versus I'm putting the head in my, you know, my head in the sand, you know, and not worrying about it, like, you know, be thoughtful about it. For, for many of us, we're going to have to move fast. Certainly in my business, which is, you know, education, online education, content development, as you said, we're moving very quickly. I don't, I don't expect Michael to be creating courses, you know, in a couple of years. You know, I really don't. Meaning it'll be AI coaches, you know, and we're releasing a product next month called Prof. AI. That will really be the way, I think almost all learning is delivered in the future will be through AI coaches and, you know, AI tutors. So I think the right way to talk it as a leader is honestly, and this is what I say to Section, and in fact, I said this on Monday at our all hands. We are adopting AI even faster at Section, both in our internal workflows and in our product roadmap. And that will do one of three things to each team. Some teams will get bigger because they get such leverage from AI and they have such an impact on the business that in fact, putting more humans into that team with the AI is good for the business. Right. You can imagine that scenario happening in all kinds of different teams in different industries, industries. Salespeople, for example. If they're more, more capable with AI, you're going to want more salespeople, likely, because they, you know, they drive revenue. Some teams will stay the same size and they'll get more done with AI depending on the team. And some teams will be smaller, I don't think dramatically smaller, but I think in some cases they will be smaller. And I'm not sure, you know, what's going to happen to each team. And that's what I. That's what, that's what you have to say as a leader. We're not sure, but let's do this together and sort of be driving that change versus being forced into that change and kind of playing catch up. So I think that's my advice to leaders, my advice to employees is similar like adopt it yourself sooner and really get comfortable and really get good at knowing where AI can help and where AI doesn't help expose AI's weaknesses. And so, you know, where it's worth working with AI and where AI is not going to Give you an assist and just be at that edge. It's just a little closer to that edge, I guess, as an individual. So you'll see what's happening again. Don't, you know, don't put your head in the sand. I think this is coming faster than we realize.
- [15:05]
Yeah, I think that's some helpful advice, and I would agree with you. I think that's how we talk about it here at User Testing as well. So putting, putting the focus a little bit on the folks listening to this podcast. So many of those individuals are, you know, people who are constantly exploring ways to elevate their customer experiences. So from your perspective, what should businesses be thinking about? Like, how would they leverage AI tools to deliver exceptional customer experiences? Do you have thoughts on that?
- [15:33]
Well, first of all, I think to step back, I think what we're seeing is a fundamental change in how customer experiences are delivered. And I think the best example of that for me as a consumer, frankly, is travel planning. You know, we all know how to plan a trip when we're online using a browser. Right. We have a lot of tabs open. We've got TripAdvisor and Expedia and, you know, all kinds of websites. And we're kind of, you know, just spending probably an hour or two getting to a decision around a weekend trip or something like that. And when you do that with AI, we don't yet trust AI to the extent we need to, to probably make, you know, follow all the recommendations of AI. But I think we're getting close. When you do that with AI, plan your own sort of vacation, you realize how frictionless the experience is when compared to the old way, because you've got one conversation with one ui, you know, with one source of knowledge or, you know, one expert and that and, and that really, to me, for all of us who are building consumer experiences, that to me is, is, I think, the most dramatic change or paradigm shift for, for consumer experiences. Because of generative AI, which is interface, friction will not be tolerated. Consumers, particularly younger. Right. Consumers, AI, the AI native generation, they're growing up in a world where they won't tolerate that kind of friction. They won't think about. Look at the Amazon product page. Talk about a page full of friction. Right. There's the product.
- [17:04]
Sure.
- [17:04]
Somewhere in there is the product, and there's ads and they're sponsored and there's all kinds of information on that page. And a lot of cognitive load is therefore pushed to the consumer when they reach that page to make a decision. I just think that Again, in a few years, we'll look back at that product page on Amazon's website and say what we were thinking. That's just loaded with friction and loaded with cognitive burden for the consumer. And AI can really wipe that away. And I get it. I can see why Amazon's not in any rush to build that page. I can see why Google's been rolling out these changes slowly to search results and so on, because it really threaten their business models, as you might expect. But is a much better consumer experience around the corner when it's generated by AI?
- [17:54]
Yeah. So your example of vacation planning reminds me of some of the demos that we've seen from Anthropic's computer use model, OpenAI's operator model. That's. Those are kind of like the truly autonomous agentic AI models that go out and do the things for you instead of you having to do them. So not only is that reducing the friction of you having to sift through all the different travel websites, it's just doing it for you. Right. There's like, ideally no friction there. Do you have any thoughts of what should businesses be thinking about right now? If we are actually going to quickly enter into a world where autonomous agents are doing a lot of the things that we as humans do today?
- [18:37]
I just think you got to stand up a lot of pilots. I think you need to be prototyping like crazy. And. And I don't mean toys, you know, meaning. I don't mean like the AI, AI, you know, powered kiosk for Burger King or whoever it is. Like, forget the AI toys. It's just. That's just noise, press releases and a waste of time. I think what businesses ought to be doing is going in two places internally, in their internal workflows and in their customer experiences. Whether they sell a product or a service, I don't think it matters. And anything you deliver to customers and, and just, you know, pick off what look like good fits for an AI prototype and expose that prototype to either consumers or internal employees and just learn, learn fast. Right? And know that I think this is a lot like drilling for oil. Maybe I finished watching Landman, so I've got oil in my brain. But you're gonna. A lot of. It's gonna be a lot of dry wells, a lot of empty holes, because that's where AI is right now. A lot of agents won't work or they're going to crap out when they try to do the transaction. That's okay. That's not a reason to dismiss AI. I think A lot of people, Michael, either consumers or companies are bouncing off AI because they do have these false starts and they do have these dry holes and they do bounce off and they're like, oh, it's not ready for prime time. That's a mistake. I think some of this is ready for prime time and some of it is not. And I think you need that kind of resilience and that sort of experiment mindset just to get going. And some of this is going to work, some of it won't. But. But I wouldn't. I would not stand still or, you know.
- [20:12]
Yeah, I think the oil drilling example, that metaphor, really, that. That feels like it hits home for me. I agree that not all the experiments are going to work, but those that do are going to produce enough value to make the ones that didn't work worth it. Right. So one of the things, Greg, that you and I have talked about is who. Who within an organization should be driving this mission? I think that no organization has kind of like figured this out really well, or maybe, maybe a few have. But enter. What are your thoughts? Like, if an organization comes to you and says, like, we're serious about AI transformation, we want to do this in our products, we want to do this in our business, who should we get to drive this for us? What would your answer be?
- [20:57]
Yeah, my answer would be it's probably three people combined into one. If you could do that, and if you can't, then, you know, pick one. The other two support, and the three are the technologist. You need someone to buy the AI, you know, connect it to the data sources, the company data, secure it all, and get it in the hands of employees. So that's the technology cto, head of engineering, whoever owns that responsibility inside the org, the change manager, who. Who is the person that is best at not just the upskilling, but the change management required. Because it's really about change management. It's about getting humans less anxious and more confident and frankly, more willing to accept and work with the AI. You know, some of that. Some of this latest research about employees revealed that some employees are actively resisting. You know, they're fighting back because. Because their perception is, you know, the AI is coming for their jobs. It kind of reminds me of, you know, people in San Francisco putting cones on the top of Waymos. So the waymos get confused and can't drive. Right. If employees start doing that, that's not a good thing, right? AI is not gonna. Not gonna land inside your organization successfully. So you need the change manager and the third really is the business person and the business person who understands the workflows, understands how work is done and, or how products get built, depending on the kind of company you are. Right. So you know, understands where AI is going to basically insert itself successfully in those workflows or product roadmaps. I think that's the lead person. This is not a technology challenge primarily. I think there sure are some technology challenges, but I think the primary challenge here is knowing where to insert AI in the right places and then getting the humans to support that effort. When you think about deploying an LLM specifically, I think agents are different. When you think about deploying AN LLM like ChatGPT or Microsoft Copilot, it's only as good as the humans that work with it. So if they're resisting, if they're anxious, if they're not willing to learn, then the LLM will not deliver the output and therefore the productivity gain, the ROI that clients need. I think that's why a lot of companies will struggle with their LLM deployments is that they just, just haven't done the change management, they haven't done the upskilling, they haven't really got the humans ready for working with LLM. I do think some companies will skip that phase and try to go just to agents because when you go to agents, you're essentially trying to take the human out of the loop. Right. And put them maybe at the beginning or the end, but not in the middle. When you deploy an LLM, you are putting an LLM in the hands of every employee. I would assume that that's my recommendation. Give every employee access to a full featured LLM that can access company data. Maybe not all the data, but can access you enough of the company data to be useful. I think every organization should do that. Some will not and or some will struggle doing that because of that Anxiety resistance, lack of training, lack of change management. And again, agents I think do offer understandably, you know, some upside around. Can, can we take the human out of that loop and just really make it frictionless in terms of the workflow and get the output we want. So, you know, we'll see both.
- [24:04]
Yeah, I think the three Personas that you describe that I would agree with that 100%. I think a skill that I see a lot of businesses under indexing on when choosing who is the right individual or group of individuals to help drive this. So that skill that they're under indexing on as I think is creativity. This requires creative application to figure out where does this fit into all of the different roles, all of the different, you know, functions within the business. And I think the reason why that's so important is because using LLMs, deploying LLMs is a very. LLMs are just very abstract, right? Like, we don't know fully how they work, we don't know fully where they apply. So we need someone, you know, a creative individual can help translate that abstraction layer into something that feels real and concrete for a lot of different Personas within the business. That probably fits into, like that transformation person that you described. You know, whether you agree or disagree with that, like.
- [25:08]
Yeah, I do. I mean, I agree. Meaning you want someone to really can span the business. It's one of the, it's the Achilles heel and superpower of AI. The superpower of AI, generative AI specifically, and something like a chatbot, like ChatGPT, is that it can do a lot of things. It doesn't sit in one function. Right. It doesn't just help the marketers. You know, it helps pretty much everybody potentially in the organization. So you do need someone that does have that, that sort of. But critical and creative thinking around, you know, where should we best deploy the AI and in what kind of use cases. Right. And that's sort of its Achilles heel because we're used to software that does one thing really well, and we're used to deploying solutions that go into a specific workflow or specific function. And this thing, you know, AI, generative AI just kind of goes everywhere and frankly does some things really well and other things not so not great. And so it's again, it's sort of the, it's like a lot. It's like humans in this respect, right? Our, our superpowers, when taken to the extreme, are our weaknesses. They're our Achilles heel. And I think this is the case with AI as well.
- [26:13]
Okay, so aside from the obvious LLMs, chat, GPT, Claude, Gemini, what is. What's an interesting AI tool that you're working with that it's provided you with a ton of value?
- [26:27]
Yeah, you just named them, really. I used to play around more with writing AI or image generation AI more for fun or for party tricks. Frankly, I can't keep up with the capabilities that are in the base models, meaning Perplexity, chatgpt and Anthropics, Claude. So I'm more interested in frankly mastering projects and mastering no code bots like GPTs or even operator functions. So I, I'm con. I've got a day job, Michael, like you. And so I, you know, I. I never have as much time as I want or, or need to keep up. So I've decided just to frankly stay more focused on the models and exploit all the capabilities that they're releasing, because they are. They're on it. They're on a torrid pace in terms of releasing new capabilities. OpenAI's deep research being one example where, you know, I've been using that a lot the last few weeks, and I'm sure I'm not yet even, you know, kind of maximizing its value. So I got, I got enough just with those models.
- [27:31]
Okay, so what you just described is a pretty overwhelming landscape, right? Think about someone who has not quite dipped their toe into the water yet, and here's this conversation and says, all right, I really need to get started. What is that starting point?
- [27:47]
I think it's chat GPT and a few hours and take a course, you know, and email me and I'll get you a coupon for a free section course. If you're listening to this podcast. Podcast or, or we'll put it in the show notes. I. I think the right thing to do is get that AI. It's the Swiss army knife for. For AI, I think, for generative AI at ChatGPT and just get going there and, and just start to use it. If you can't use it at work for whatever reason, then use it at home and, you know, get up that learning curve. Get up that learning curve where you're really comfortable loading up PDFs and talking to AI about, you know, about these decisions or about your workflows. Just. Yeah, experiment with it. Have fun with it. You know, create some images, make a song on Suno, you know, whatever it is. That kind of floats your boat, right? Get a sense of the power and limitations of these technologies. I think you'll do that in a few hours and it'll be clear to you where it's going to add value to your life and then go there, you know, and spend your time on those use cases.
- [28:46]
Greg, that's great advice. So for links to everything that we talked about, visit usertesting.com podcast Greg, I want to thank you so much for being on the show today. I really enjoyed our conversation. I enjoy all of our conversations. So how does someone learn more about you, your thought leadership and the work that your team at doing?
- [29:08]
Yeah, I think the best place to go is Greg. Shove.com G R E G S H O V E like shove, but we say it's shove.com and everything about me is there and I spend my day job is I'm the CEO of sectionai.com which is and we're proud to have you, Michael, and User Testing as a customer. And so that's another place to find me, of course.
- [29:31]
Fantastic. Thanks again, Greg.
- [29:33]
Thank you, Michael.
- [29:34]
Michael Dominick
Want to keep the conversation going? You can find the show notes@usertesting.com podcast if you haven't already, don't forget to follow us on Apple Podcasts, Spotify, Overcast, or Google Play, so you never miss an episode. And if you enjoyed today's show, please share it with a friend or leave us a rating and review on Apple Podcasts. And until next time, this is Insights Unlocked, an original podcast from User Testing.