You can’t get value from generative AI without good data, and that’s why author and venture capitalist Geoffrey Moore dubs the current moment “digital transformation, Act 2.”
“You have to feed the beast, [and] it needs a huge amount of data,” says Moore, the author of best-selling books such as Crossing the Chasm, Zone to Win, and his most recent, The Infinite Staircase: What the Universe Tells Us About Life Ethics and Mortality.
Our conversation delved into Moore’s insights into how his “Zone” model can apply to generative AI; the Apple VisionPro headset and whether virtual reality is finally crossing the chasm; and the psychology behind recent innovation cutbacks.
You can listen to this episode of our Innovation Answered podcast by clicking play above, or subscribing to it on your favorite podcasting platform. If you have an Alexa device, just say “Alexa, play the Innovation Answered podcast.”
Scott Kirsner:
Welcome. You’re listening to the Innovation Answered podcast. Innovation Answered is the podcast from InnoLead, the web’s most useful resource for corporate innovators and change-makers. If that sounds like you, we encourage you to subscribe to the podcast, so you’ll catch all of our future episodes.
I’m Scott Kirsner, CEO and Co-Founder of InnoLead, and in this episode, we catch up with Geoffrey Moore, the author, consultant, and venture capitalist.
If you’ve ever talked about a technology “crossing the chasm” from the realm of early adopters into the early majority, or your company has put a high-potential project into “the incubation zone,” you’ve been using language given to us by Moore. His best-selling books include Zone to Win, Inside the Tornado, and Crossing the Chasm — and lately, he’s been publishing a series of insightful LinkedIn articles on topics including the future of the office, rethinking personalization, and why it’s time to stop talking about generative AI and start doing something with it.
I spoke to Moore in February 2024.
• • •
Geoffrey, models come and models go. If you’re a consultant, you pretty much have to invent a new model with every LinkedIn post or webinar these days. But we talk with a lot of executives at large companies, and I have to say, for a book that’s a little over seven years old, I’m amazed how many people say they’re using your Zone to Win model to help them think about managing innovative projects and initiatives. Could you lay out the four zones for us?
Geoffrey Moore:
Sure, and what we’re trying to do here is say how would an established enterprise both pursue its core business in a relatively mature category, and simultaneously engage with next-generation market development dynamics, which are very unpredictable, particularly if you’re a publicly-held company, and so you’re under the pressure of earnings of performance, every quarter, etc. The Zone model was written with Marc Benioff’s team at Salesforce and Satya Nadella’s team at Microsoft, so two really good teams. I think part of the reason it has staying in power is frankly, you stand on the shoulders of giants.
The first zone, we just call it the performance zone, that’s the zone of your core business. It’s the stuff you report out, it’s how you deliver value to the world. It’s how you fulfill your mission. In that sense, it’s ultimately the zone that matters, because that’s the zone that you act in the world through. Also, we tend to measure it financially, which has good and bad components to it, but it’s deeply ingrained with financial metrics. All the earnings calls are typically heavily focused on the financial performance of the performance zone.
Now, you can’t run a performance zone without what we call the productivity zone. That’s all the functions behind the scenes that make it possible for you to operate at scale in front of the audience. It’s all finance, HR, marketing, customer support, facilities, legal, you name it. There’s usually twice as many people in the productivity zone as there are in the performance zone. In total, it’s 90 percent plus of your operating budget in those two zones, and that’s fine.
The key lesson of [Zone to Win] on the incubation zone is to run it as if you were a venture capitalist, not as if you were a member of the productivity zone, which is where where it’s normally run.
Now, disruptive innovation comes along, we’re going to talk maybe a little bit later about Gen AI — it’s the latest, but we’ve had cloud, we’ve had mobile, we’ve had everything and its mother. So you have a perfectly good business. What are you supposed to do with this new thing? We talk about an incubation zone. Most corporations had a skunkworks, or they had something that was like that, but they were not managing it well at all. The key lesson of the book on the incubation zone is to run it as if you were a venture capitalist, not as if you were a member of the productivity zone, which is where where it’s normally run. When process models from the productivity zone are applied to incubation, it just doesn’t work. But the VCs know how to operate in this model. You don’t have a VC financial model, but you want a VC operating model. The key idea is you fund entrepreneurs, you fund to milestones, until you’ve done something meaningful in the market, nothing real has really happened. You kind of work your way up from there.
That’s a discipline that is still a challenging discipline to implement inside a public company. But at least there’s a very clear roadmap of what we would like to do. Then, the fourth zone, which is the only zone that you would say is normally actually inactive, is called the transformation zone. That’s where you go, “OK, we have this next-generation innovation that is now at a material stage.” It’s typically what we would call “inside the tornado,” meaning it’s changing the world now. We’re either going to get on this bus or miss the bus. And that is going to cause you to in some way, put at risk or take down your performance contract, because you’re gonna put a bunch of resources into the new thing. And they’re not going to be used to keep the old thing optimized and it’s going to show. That was where public companies just struggled over and over again…. The idea behind that is two things. One, transformation is horrible. People like to talk about transformation like I’m a caterpillar that has to become a butterfly. No, you’re not. It’s going to be awful. The risks are horrible, and you’re going to have to explain yourself to a whole bunch of people that aren’t gonna believe you. You don’t want to do transformations often. But when you need to, you need to, and so how would you? The key idea behind that zone is, once you start a transformation, you cannot take your foot off the accelerator. You just must see it all the way through even though it’s gonna get dark before it gets light.
Scott Kirsner:
You wrote a piece at the beginning of 2024 where you said, “Hey, this needs to be the year, if we’re talking about generative AI and the potential there, this needs to be the year that companies move from talking about generative AI to actually doing something about it.” I would say that a lot of companies spent last year not only talking about it, but probably writing policies about how these are the guardrails, here’s what you can’t do, here’s what you can’t upload. How would you use the zone framework to nudge companies into action? Or how would I use it if I’m an employee in a large company? How might I use the zone framework to say, “OK, let’s get things in gear?”
Geoffrey Moore:
The first key observation that it’s implicit in what you just said, in the post I made, is you can use Gen AI in any of the four zones, but don’t use it in all four zones at the same time. The most obvious place that we’re going to use it is in the incubation zone. So incorporated Gen AI. If you’re doing that, though, you better make your Gen AI distinctively differentiated. The incubation zone is for catching the next wave and participating excitingly. If you’re just playing with Gen AI in that zone, it’s gonna be corporate entertainment, and nothing meaningful is gonna happen. Milestone-based funding is great, because it says, “OK, can you use Gen AI to go get a customer that nobody could have ever gotten before?” That’s amazing.
Having said that, I think where more people are going to use Gen AI in 2024, is in their productivity zone. Because it’s obvious that there’s a bunch of just stupid stuff you have to get done, and Gen AI is a very useful tool for getting people through that. Getting people through self-service things on your web, creating first drafts of marketing texts, or first drafts of user manuals, or any of that stuff. Let’s just use that to take some work off of our plates. Then we can spend time editing it, and adding more value. That’s in the productivity zone. The key thing there is you’ve got to make sure you capture learnings. It’s not enough to just use it there, you’ve got to be able to get back to the corporation and say, I now know something about Gen AI that we didn’t know before. It’s a two-part mandate; become more productive, but also bring back new learnings for us.
The key thing with Gen AI is that you can’t come out of 2024 knowing as much about it as you do today because frankly, you don’t know squat about it today.
Frankly, it’s not a good use of our time. If you can get Gen AI and your performance-owned products in a way that is not disruptive, that doesn’t cause your existing customers grief, and that in fact, adds relief rather than grief, way to go. Salesforce has, I think, a very good job at taking a first-generation Gen AI, and overlaying it into sales, cloud service, marketing cloud, and all the various clouds. That’s great, but there now you want to be almost invisible. You’re just adding, but again, as with the other two zones, you must learn.
The key thing with Gen AI is that you can’t come out of 2024 knowing as much about it as you do today because frankly, you don’t know squat about it today. So at the end of the year, you’ll say, “Look, we’ve done something material with this thing, and here’s what worked, what didn’t work, and here’s how we think we’re smarter about 2025.” That notion that you would use Gen AI and the transformation zone seems very improbable to me in 2024. It’s too new and transformations are just too risky. I suppose there could be a corner case, but that didn’t come to mind.
Scott Kirsner:
I want to just diverge for a second and ask about the leadership of those experiments because when you talk about the productivity zone, I think IT typically fits in that zone, right? There in the back office, they’re supporting everybody. But you often see IT leaders now in the driver’s seat when it comes to assessing the landscape of Gen AI tools, assessing the landscape, and deciding what vendors we’re going to buy from and who are we going to bring in. How are we going to give people access? I wonder if that is the right person to be leading here? Are you seeing some examples where it’s not the CIO in the driver’s seat, but you might have a chief operating officer or chief marketing officer, or someone different, who is helping make sure that, there are some strategic decisions around where which one of these zones we’re going to use Gen AI on?
Geoffrey Moore:
Venture capital, increasingly over the last 10 or 15 years, has invested in what I call a two-headed entrepreneur. There’s a technical leader and there’s a business leader, and you need them to be joined at the hip. That’s what you need here. Obviously, Gen AI has a bunch of technical stuff that only somebody from IT is going to get right. By the way, there’s a technology enthusiast in there who’s just dying to get at it, but if you let them lead, the problem is that they have the shiny object attraction issues, which is great and you want them to, but you need the other head now.
I think the other head has to be a business leader who, first of all, is willing to accept the assignment that I am learning on behalf of the entire corporation and I have a responsibility to communicate that learning back to the corporation. It has to be high-quality reviewed learning, not just, “Hey, this is what I think I figured out so that person has got to take whatever task.” You put generative AI on top and take that task very seriously, maybe more seriously than you normally would because you want to say, “Did we make a material difference in the execution of this body of work?”
That requires real thoughtfulness real engagement, you can’t just fire and forget. I have the experience of watching people do a lot of fire and forget with the next-generation stuff. My phrase for that is corporate entertainment. It creates great demos, it’s kind of interesting, we learned something, and we have a great conversation, but at the end of the day, we didn’t make any progress.
Scott Kirsner:
You talk about this idea of return on innovation, which is not about entertainment, and just running experiments for fun. But for you, that’s the difference between just being innovative, and looking innovative, versus actually delivering things that are going to differentiate you or reduce costs or close a competitive gap.
So, in 2022, and 2023, we saw a lot of companies cutting back on their investment in innovation and R&D, and cutting headcount in those teams. Often it was amidst the talk of “Are we in a recession? This seems like a recession, the economists all agree there’s a recession,” and it turns out, it wasn’t. We still seem to be in this efficiency-streamlining mindset at the moment. What do you think changes about that and are we just at a time when companies have retreated into their shell a little bit, retreated into the performance zone and the core business, and it’s going to take some dynamic to get them back into the like, lean forward, let’s incubate, let’s experiment more?
I think until the elections in November, people are still going to hedge their bets and stay back from a growth mindset.
Geoffrey Moore:
You and I have been in the tech industry our whole life, so growth mindset was baked into the industry from birth. I think what’s weird about 2023 is that we pulled back from the growth mindset and I think 2024 is not going to be any different. First of all, two things have happened which I don’t think we’ve seen before. One is geopolitical risk, which has never really been part of the radar of the tech-enabled industries. Well, this year it is — there is political risk inside our country, and geopolitical issues outside of our country, and that’s causing people to hesitate on investment. I think until the elections in November, people are still going to hedge their bets and stay back from a growth mindset.
So, you say, “Well, how bad is that?“ Well, two things would make it bad. One is, while you’re hanging back, there is a tornado that has formed. An entire generation of companies obliterate your industry because you’re hanging back. I think of the EV because EV is pulling back. If you’re General Motors, you think I’ve got another year but you could say just think of that as so-so. If you want to be a first mover, the problem is now you need venture money. Venture money may come back, but venture money still feels a little bit battered from the last couple of years.
Most disruptive innovations come from large companies needing to catch up to disrupt doors. As long as there are no disruptors at the door, you can postpone this. That’s where you can cut back your incubation zone because its primary function is to keep you relevant, as these waves come along. If the waves are not there, I don’t want to spend that money this year. This notion that you should spend a lot on R&D every year if you’re an established enterprise, that’s not quite the right idea. You would spend a lot and continuously sustain R&D for your core business. You probably should pay a little bit of attention to the next new thing because you don’t want to be blindsided. You don’t put a lot of money into that until you’re going, “OMG, we better get on this one.”
Scott Kirsner:
I had been thinking from the 2013-ish period, the year that the term unicorn was coined by Aileen Lee, from that period to the launch of ChatGPT in November of 2022, the fear of the unicorn, was where disruption was coming. It was like, “Stripe was coming for you, financial services, and Airbnb was coming for you, hotels.” My current belief is maybe generative AI is what causes the concern and fear but I don’t know if that’s going to fuel the creation of innovation teams or hiring R&D people. It’s just fueling some strategic what do we build versus what do we buy? How do we hire people?
Geoffrey Moore:
I think what it’s actually going to do is get rid of the bottom third of your white-collar workforce. If you think about white-collar workers, there’s amazing stuff that has to be done. That’s people, I’m sorry. There’s a bunch of stuff that’s mission-critical. You can get a little help, but it’s still people. There’s a whole lot of what I would just call stupid stuff. You have to do it. Its compliance-oriented, and compliance-oriented responsibilities are ideally suited — particularly modestly risk-bearing, and low risk-bearing compliance activities — are where this stuff is going to work…. That employs a whole lot of people.
There’s a ton of people that are employed, onshore and offshore. That’s going to come with the return on innovation. Remember, you can differentiate and get ahead. That’s not what we’re talking about. You can neutralize or catch up. That’s not what we’re talking about. You can optimize. That is what we’re talking about. It won’t look like innovation necessarily to the outside, it’ll look like innovation to the person who lost their job. It will look like innovation is somebody who inspects their quarterly P&L, because you’re going to optimize the P&L with this stuff.
Eventually, we’re going to free up talent to take on whatever the next big thing is. People are always worried that they’re going to lose all these jobs in the short-term. That’s true and it’s very painful and it’s worth attending to, but the world is not going to run out of work. We’re not gonna run out of problems to solve. Just look around, I don’t think we’re out of problems yet.
Scott Kirsner:
Yeah, but it’s interesting with AI when, I know you’re a fan of good metaphors, it’s almost like when you think about the moment of electrification right? When businesses had to shift from, you know, whatever. We’ve got oil lamps and candles and chasing whales. The era of electrification was the focus for companies.
I might call AI digital transformation, Act 2. AI doesn’t work unless you’ve done digital transformation in the first place. You have to feed the beast, it needs a huge amount of data…
Geoffrey Moore:
In that sense, I might call AI digital transformation, Act 2. AI doesn’t work unless you’ve done digital transformation in the first place. You have to feed the beast, [and] it needs a huge amount of data… Implied in AI is you have digitally transformed. Well, not everybody has. Part of the reason it’s gonna take longer is that it just is going to take longer. The kind of digital transformation you need is a bunch of heavy lifting that you have to get done in that heavy-lifting moment. You know, you’ve got to, you’ve got to sort of say, okay, it takes time. I do agree with you that we are going to rethink the distribution of knowledge work.
Scott Kirsner:
I can’t wrap up without asking you one Crossing the Chasm question. We just watched the launch of Apple’s Vision Pro AR and VR headset at $3500. If the world is divided into innovators, and early adopters, early majority, late majority, and laggards, and the chasm is between the early adopters and early majority, where are we with AR and VR, which I like to point out, academics like Ivan Sutherland began working on in the 1960s.
Geoffrey Moore:
We had VR for 30 years. We’re on the early market side of the chasm and I’m not sure we can see the other side yet. We’ve talked about this a lot. We kept on saying, “Well, we’re gonna have augmented reality, and field technicians are going to look at the machine and they’re going be on it and God bless.” We’re nowhere near that and, by the way, I think Apple is a perfect early-market product.
They’ve sold 180,000 of these things at $3000-plus. Are you kidding me? Now, that is the Apple brand and a bunch of people who have $3500 to buy toys with, because it’s a toy. Now is it a toy in the hands of a genius? It’s a piano. I’m sure there [are] geniuses who will do amazing things with this thing. But would a pragmatic person buy this now? Not in their lifetime. Let me say it doesn’t qualify for any of the pragmatist buying criteria. Are my friends using it yet? Does it solve a problem that I can’t solve without it? No.
Scott Kirsner:
Since we’re talking about crossing the chasm, what’s your view on how generative AI crossed the chasm so quickly, and so easily and seemingly with no marketing campaign or no marketing spend?
Geoffrey Moore:
We have been talking about AI and machine learning for at least a decade but it was the AI and machine learning that happened more behind the scenes, typically around data analytics, and then increasingly around real-time data analytics. Then, there was a focus on real-time analytics and operational real-time transactional changes in commerce and advertising placement, and increasingly around preventative maintenance. But all of that requires a lot of heavy lifting. Then, all of a sudden, in November of 2020, Chat GPT 3.5 came out, and you go, “What? Is that possible?”
There’s no chasm [with generative AI.] It doesn’t cost anything to adopt it. I don’t have to learn how to use it. It just works.
There’s no chasm. It doesn’t cost anything to adopt it. I don’t have to learn how to use it. It just works. I can integrate it into my way [of working] if I want to. If I’m a student at Harvard, I can submit it as my essay. It really wasn’t that disruptive.
Now what are we learning about it subsequently? Gen AI, the way it works, will give you the most plausible, probable response to any questions. If you’re looking for something to just get oriented, it’s fabulous. If you’re looking to contribute it as a contribution to human knowledge, it’s useless. It’s just summarizing what we already know. But for a whole lot of things in life, summarizing what we already know is okay.
It’s usable. And the fact that it was so psychologically available.. . We talk about low code and no code, and this was nothing. It just answered your question. It’s pretty exciting and, therefore, I think highly disruptive, eventually.
The question really is, how do we absorb this? You mentioned already all the compliance and regulatory concerns, and by the way, our elections are going to make this painfully obvious. There are still some hurdles to cross. But again, if there’s a chasm, it’s not across it. If you say crossing the chasm it hasn’t been institutionalized anywhere yet. We’re going to get close, it’s going to get institutionalized with chatbots and customer support.
The quality of human support is not particularly distinguished. Gen AI probably does a better job, in general, answering level one customer support questions, maybe even level two, than people do. So, I think that’s going to happen pretty darn quickly.
Scott Kirsner:
Geoff, it has been so great having you on the Innovation Answered podcast. I really appreciate you taking the time and so enjoyed hearing your thoughts.
Geoffrey Moore:
Well, Scott, I enjoyed it very much. I always enjoy talking with you, take care.
Scott Kirsner:
Geoffrey Moore’s latest book is The Infinite Staircase: What the Universe Tells Us About Life Ethics and Mortality. You can follow him on LinkedIn or learn more about his work at GeoffreyAMoore.com You can find out more about InnoLead, sign up for our email newsletter, or find out about our 2024 in-person conferences in Silicon Valley and Boston, at InnoLead.com. To listen to more than 60 of our earlier podcast episodes, search for Innovation Answered on your podcast platform of choice. Thanks to Geoffrey Moore and recording engineer David Swope for his help on this episode.
