In today's book our guest highlights how artificial intelligence is becoming less artificial and more intelligent, how instead of data hungry approaches to AI innovators are pursuing data efficient approaches that enable machines to learn as humans do and how, instead of replacing workers with machines they are unleashing human expertise to create, human centered AI. lEading companies use these basic building blocks of business. Intelligence, data, experience, architecture and strategy.
The IDEAS framework to transform competition, optimize post pandemic approaches to work and talent and show the way towards a sustainable future.It is s a great pleasure. I've been waiting for quite some time. The books are on the shelf there behind me to welcome senior technology advisor to Accenture, former CTO and head of innovation for Accenture, the author of radically human and human machine and the human machine updated version paul daugherty welcome to the show.
It's a pleasure to be here, Aiden. Really looking forward to the conversation. It's great to have you man and as an irish man i want to say also paul is the science foundation ireland saint patrick's day science medal holder as well Yeah, that was an amazing event to receive that from the Taoiseach a couple years ago and yeah, that's another one of my connection points to Ireland, which is another reason why it's great to talk to you.
was great to have you man i really love the book i've shared it with so many people because you really for such an expert in the field of technology and AI in particular. You do really make it accessible. There was stuff that you just made so accessible for me. And I read about this stuff quite a lot, but I thought we'd start off by sharing the context for both these books, because human and machine came first and then radically human.
And then you updated human and machine because the world changed dramatically in between the additions and between starting human and machine and radically human. So let's share a bit of context for that. And then we'll share some of the changes that you've seen and some of the context for. All the work that you do the books and really all the work that I've done ties back to, experience back in my university days, a few decades, several decades ago now.
Where I got into, , AI and also cognitive psychology and some fields which we can talk about a bit more, but it was really formative experience that really set me on this path of understanding the impact that technology has on people and how people use technology. And that's been at the heart of everything I've done over my career.
And that's what led to, , human plus machine, radically human and , all the work that I did with the fundamental premise that, , technology is an amazing tool that allows people to be better. And I don't think AI is any different than any technology that came before. So, what we're really focusing on with the research and work we're doing is how do we use this technology to help us as people, , work, live and play, more effectively, than we could before. So that's what set us on the course.
I think what are the points , you were mentioning in the intro there was is that there was an interesting point between the first book and the second book which, Which is in the first book, we talked about some research where companies that were the leaders in applying digital technologies, including A. I., were outperforming the rest by a factor of two X in terms of revenue growth and other metrics, and that was impressive in and of its own right. Enough of a reason.
, to put this on the agenda, if you're a business or organization leader. But then what was interesting is after the first book, , we did some new research, and this is talking to thousands of companies, 8, 000 companies in the first iteration of this study, over 4, 000 in the second iteration of the study. I think it's one of the largest ever, maybe the largest ever study like this done of the use of technology and businesses around the world.
And we were really surprised the second time, this was after COVID, and we found , that those leaders. We're just achieving a 2x performance advantage. It was 5x, it accelerated. Leaders were accelerating, and that was really the insight that led us to write the second book, Radically Human, which talked about the need for every organization now to, to get, , to figure out what this meant to them and how they, , how they , could get an advantage.
And we covered a second category of companies called leap progress, where we're using the technology to, , to go from, being laggards to change the game in their industry. So, and that then led to, with generative AI and everything else that came that led to us writing the third book, which was the updated version of human plus machine, because we believe the the premise of that first book are more true than ever. in the era of generative AI that we're moving into.
i'd love to share in a moment tech vision twenty twenty four the work that you've done before that i was so interested by the fact that the pandemic was an accelerant for all this stuff. Most people in technology had been talking about for quite some time telling people you need to build capability inside your organization you need to understand this stuff and you're a little bit like chicken little telling people the skies. falling down and then the pandemic comes and accelerates everything.
And some of the stats you mentioned in the book and some of the senior leaders that you talk to were like, we were trying to get this through something that would have taken years before happened in two weeks. And I thought that was really fascinating because it pushed businesses to become technology businesses, which was going to happen anyway, but the accelerant of the pandemic changed everything. Yeah, no, and there was an interesting dynamic.
Yeah, it was definitely an accelerant, but there was also another thing that the pandemic did , was , when the tide went out, so to speak, it exposed, , the shaky foundation that many organizations had. Everybody had been talking about digital and building digital and every, all that sort of stuff. But when you were forced to create digital supply chains, when you were forced to do Touchless contactless payments.
When you're forced to allow all your employees to work from home, companies realize, Hey, maybe I'm not as digital as I thought it was. And that's what led to the acceleration. So is that realization that we're not really, , where we need to be? It's like to extend, , metaphor you were using earlier. It's like the emperor, , had no clothes in some respects as companies looked at it.
And so we saw, Yeah. That coming out of that period, , there was a 63 percent increase in technology, meaning, 63 percent t of companies that weren't previously using some of these technologies adopted it very quickly. And that's the kind of acceleration that we saw.
Before I get stuck into this book in particular, radically human, which I absolutely love both of them, but I wanted to share Tech Vision 2024 before we get outta 2024 , you consistently update this, but let's share a little bit about that work and then I'd get stuck into the book. We doo this technology vision every year. I've been doing it for well over a decade.
And it's been a remarkably accurate forecast of, , the trends that will impact business and that business leaders need to think about. And 2024 was a particularly important year, I think, from a number of perspectives.
And I think before I talk about what the trends are, you have to back up and realize that this is a really important moment, an epochal moment in human history, I believe, in terms of what we're experiencing with , generative AI isn't just another technology, it is, , simply put the most powerful digital technology of the 70 year computerization era that we've been in because it has this more human like characteristic. And it's already started to make massive changes.
I think most of your listeners, I'm sure, have played around, experienced written a Shakespearean sonnet or something with chat GPT. So you see the power and the power of this to transform business, transform organizations, transform the way things work is really stunning. And I think it's what I can, , it's one of the four, I can come back to this. One of the four points that really surprised me in my , four decade career with technology. And so we just have to put that in context.
And then when we talked about the vision for this year, we singled out a few trends that I think are really important to tune into. One of them is, you know, AI, and we talk about, a match made in AI is clever phrase we use, but it's talking about this human plus machine nature of AI now supercharged with a human like capabilities of generative AI. And we'll talk a lot about more about this during the podcast, I'm sure. But that was the first, the first trend.
And we've all seen this one firsthand. The second trend we talked about was something we called Meet My Agent, which is talking about how AI would become embedded , in agent personas so that you could interact with them. Not just prompt and ask a question, but you interact with, in my case, I can interact with a coding agent that can help me code or a testing agent or a research agent that can look for information.
And this rise of agentic computing is really what stands to transform the way businesses work, more than anything else. The third trend we , talk about is the space we need, talking about how digital will extend itself to three dimensions. We're stuck up where I'm looking at you in the screen now, and your viewers are looking at these two dimensional screens.
And it's kind of crazy to me that that we live in a three dimensional world, but , we've just accepted the fact that , We can only do , digital interaction in two dimensions. That's changing. Now, we talked a lot about the metaverse from previous years. Apple released their headset , not too long ago. Lots of things moving in this space. Meta has the new, glasses.
Watch this space because it's inevitable and it'll happen faster than we think that we'll be interacting more in three dimensions than two dimensions. Digitally. And then the final trend we talk about is how this starts to, , move to the edge and impact everything we do.
And we call this trend the body's electronic or the body electronic talking about how devices and instrumentation and awareness of every move we make and ubiquitous computing the environment around us really has some risks we need to manage for obvious reasons, but also stands to really add a lot of opportunity for us. , I wanted to come back to some of the human challenges , and by this I mean in organizations so. We work with senior leaders all over the world.
Many of the show listeners are senior leaders, CTOs in many cases, and they're frustrated oftentimes that their organizations won't move until there's clear evidence that there's need to move. And one of the things you said absolutely resonated with me, which was, you said a large majority of companies used. Technologies as a lifeline, not as engines of innovation. So this was during the pandemic and you said, like so many companies that innovate under duress, not by choice.
And this is just such a frustration for so many of our listeners. They're the change makers that are heads of innovation. They're the people who are crying. We need to make this change. And I wondered, what did you see as the magic ingredients that would help people see that future? I know. A lot of the work you do is this show them glimpses of that future but still people get stuck in the status quo instead of embracing these technologies early. There's a variety of thoughts on that one.
You know, the the phrase never waste a good crisis comes to mind. And that's, I think what, a number of, leaders used to their advantage back in that period, , COVID and the period following it, I, there was one CTO of a large manufacturing organization that I talked to and worked with. And they had, , visions and plans of using , virtual reality, augmented reality to do inspections of their plants globally.
So they didn't have to send experts flying around the world, could never get it off the ground, could never get adoption until COVID happened. And then, by the way, they had a PO, a proof of concept. They had a minimum viable product within a couple of weeks.
They got this new thing up and running and we're using this new virtual augmented reality approach to do things and and it it took off and became the viable natural way to do things and another, , another technology leader, , , told me, , that COVID , had given them the business case now for innovation because people were willing to break through innovation. Think about, touchless payments and things like that, that that accelerated during the pandemic. So that's that's one thing.
The other thing I think is , in the new book in the subtitle we talk about , reimagining work in the age of ai and that word reimagining we chose very intentionally. , some of the people listening might be chief Information Officers at companies that would say, you have to think about that. I not just being, I think Chief info the, I like underplays, what so many CIOs do. Chief Information Officer. I think it's really a Chief Innovation Officer. It's the Chief Imagination Officer.
'cause I think the technology leaders, whether you're CIO or CTO or whatever role you might have. need to be the evangelist and bring the imagination across the organization so that they can help the other leaders, help the C suite imagine where to go with the technology. And we're seeing that happen now with generative AI. There's been a shift when, chat GPT is less than two years old now, or it'll be just about two years old when many every viewers are watching this. And And it shifted.
A lot of the initial, we did our surveys, all the research we do. Initially, people were saying, Hey, this is great. I can cut these costs or I can do this process more efficiently using generative AI. But that's flipped. Now we see 60 percent of leaders looking at generative AI and the technology as a way to drive growth and improve their business drive growth, and Drive higher pricing, margins, et cetera.
And , that's an important shift because it reflects the fact that people have gotten past , the, that may be more mundane , and obvious use cases to look at how this really transformed their business in a more fundamental way. What are the things i thought of was just your own experience and your depth of research in these books and indeed the tech vision reports that you put out every year and your work with clients i think that's the difference is you work with clients as well.
So you're seeing what what clients are doing things well you're seeing what clients are laggards as well but i thought about the. The things that surprised you over the last two years, even with, since the, since chat GPT kind of made people pay attention. And I often think about visionaries and futurists and people who work in the future, even in organizations, and sometimes they're too early.
And I often think about that, that movie field of dreams with Kevin Costner is like build it and they will come and then they don't come or they're waiting, they're like, where are they?
And. There can be some innovators in organizations that that go too early and they run out of steam by saying yeah things are going to change and then there's others that just get it right and i wondered what you saw as the secret formula for being able to know when to press the accelerator when to press the clutch or the brake. Yeah, it's funny you mentioned Field of Dreams. I just rewatched that about two weeks ago. It's such a phenomenal movie and a good message, as you say.
Yeah. Timing is the, timing is the hardest thing to get right in tech, whether you're a leader, a CIO or executive investing in technology or an entrepreneur, an investor, the timing, timing is everything.
And and that's hard to get that right, which is, which is, and what I'd say with the, with the technology that we have today the phrase I use a lot is it's, it's a participant sport, not a spectator sport to really engage and understand whether the timing's right and understand, Whether you can scale into your business, you need to be, you need to be using it at, at, in some way and diving in, you can't just sit on the sidelines with this.
So it's, it's been true with many technology, but I think it's particularly true with the AI technologies of today. And I would say that applies on an individual level, not just an organization level, , for anybody listening to this, , To this right now, I challenge you.
I hope everybody's using generative AI directly, whether it be, they're using cloud or Gemini or notebook or chat, GPT, , whatever you choose perplexity use it, use, use a couple of them, because I think you'll see differences and you'll learn a lot and then figure out, in the organization, where do you get started? And , the formula that I'm seeing for, for organizations that works pretty well. , on this timing issue is think about the no regrets things that you could do easily.
Now, if you're a Microsoft Teams user, turn on, the, the team's copilot, things like that and organizations are doing that, , without those, the easy use cases applied in areas where there's clear value customer services, an area where we're seeing, Broad and really rapid experimentation and then starting, with adoption of the technology and then as you're thinking about those types of cases, I can talk a lot more about what the, table stakes
cut kind of first category is think about that second speed where you can gear it up and differentiate and drive your strategy and delight your customers and do things your competitors can't. That's the strategic level. And that might be drug discovery and life sciences. It might be capital projects, breakthroughs in the, , energy or other capital intensive industries. It might be new personal, new levels of personalization in retail and understand, can you do those?
And where are you on the timing curve for those things? And looking at those levels is what should I do? Start with now that, gear speed one, so to speak, and , when, and how do I accelerate into that strategic, , Higher gear. I was thinking about one thing that you said in the book. I love what you said there. You said as lumbering legacy, it systems struggle to keep up with new technology. Leaders found that making wise tech decisions and investments have become more difficult than ever.
Yes. You often heard from CEOs in the course of your work. We know we have to become a technology company, but what technology and I found that very important thing to call out that one, they often have the wrong it stack and it reminds me of, , here in Ireland at the moment, we're supposed to get a gig broadband.
But most people's tech can't even take that kind of speed so that most devices they have in their house can't take that speed so they're like wondering why they're not getting a gig broadband. I'm the same with many many people's tech investments and when they're very much CFO driven or EBITDA driven the investment feels oh we only did this a few years ago and now you're telling us we have to do it again i'm sure you see this a lot from CTOs.
I been a, there's been kind of a, a kind of a second punch, so to speak from that. We talked earlier about the acceleration of covid. Generative has thrown a new thing in there a new level of awareness. So in companies that are looking, saying. And even with my acceleration after COVID, boy, I'm still not where I need to be to really now add in the generative AI capabilities I need. So , it's really like the second punch or the second, , moment of awareness for a lot of organizations.
And we, we call it, I call it the digital core. Like, where are you with your digital core that runs your business? Every company is a technology company. , I've been saying that for 10 years. And. It's becoming a little bit more adopted now, which means every executive really needs to be a technology executive, be literate into some respect, and you need the right modern foundation going forward.
A digital core now, when you move into the AI era we're moving into, needs to have some new characteristics. Data becomes more important. Not just your transaction, your structured data, your data lakes and such, but the unstructured data. Have you mastered synthetic data? Do you know the synthetic data? Yeah. Meaning the data that you artificially create, but that can you can use to train your systems better or your drive your business better. Do you understand that?
, so where are you with the data agenda? Most companies are quite a bit behind. That's the first part of the digital core. The second part is your application foundation. Are your applications ready for generative AI? Some of them may be trapped in kind of legacy, old legacy systems, or you might be on old versions. From some of the enterprise software providers. Well, guess what? They're not providing generative AI on the old version.
So if you don't, if you don't upgrade to the latest, SAP, Salesforce, whatever you might be using, you can't get the generative AI. So there's the data, there's the applications and you need a new architecture. The generative AI and the new AI technology are changing the architecture, your architecture more than any technology. probably since the introduction of the PC and the network.
Do you understand those new changes, new layers, new modules, and things you need in your architecture, things like switching capability to move between different foundation models? How do you integrate different models and ensure they use data consistently to drive the right quality results for your organization? Do you have the guardrails in your architecture to enforce responsible AI? This is a technology issue, not just a, Policy of principle issue.
That's the architecture piece of data applications. Architecture less than 15 percent of organizations from the work I've done and the surveys that we've done. have that, modern digital core in place. So most have a lot of work to do. It doesn't mean you shouldn't start , with, , down the path and most are starting down the path, but you have to in parallel build that that foundation for the future.
often feel so sorry for CTOs, lots of CTOs , have a place at the boardroom table, like they deserve today, but they didn't always, but then they get this pressure, when is this going to be delivered from all different angles? And there's only a certain speed that they can go to. And I wondered, had you any advice for them? Cause I told many CTOs, I had you coming on the show and they're like, Oh, ask him about this because I get it on my ear all the time.
can we go faster and there's only a certain speed we can go and i wonder what advice you have for them Yeah, this is a, it's an interesting one. It, the answer is different for different organizations, but I think there's some, There's been two big problems for CIOs and I agree with you. I, I have many, many, many, , CIO, CTO types of friends.
I think it's one of the hardest jobs out there because of the increased pressure for innovation combined with the reality of what you have to run today, like, and keep it all going. And the two things that have been almost impossible for many technology leaders to do is build a business case to modernize, because, modernization on its own, nobody wanted to pay for unless there was some other business benefit generated.
The second thing that was hard to justify were, , just data for, data for its own sake. So, so as a result, people, , companies have under invested in their, your core applications in architecture. They've under invested in their data. I think , the only advice I'd give is that I think the fascination and interest in this new technology now is, like I said earlier with the COVID example, it's giving kind of a new lease or new business, a new life to business cases around this.
So, how do you tie in some of the modernization of Along with the value you're driving as you as you implement some of the new capability, for example, if you're, implementing some new customer service agent related capability into the into your front office, which many, many organizations are, I can give you some examples in a minute. Then how do you upgrade some of the data and the data and capability at the core of it, which we're seeing organizations be able to do.
So that's one thing , is it's giving it a new way to package, , how you drive the innovation and business case to, to tackle some of these would have been intractable problems around modernization and data. The second is that. The new technology provides you with new techniques to use. One thing we're seeing that's interesting in the mainframe area, as an example I'm sure some of the listeners have mainframe systems still out there.
There's still a lot, a ton of mainframe systems out there driving a lot of value for organizations. Well, what, one of the One of the thoughts people have with GenreVI is, well, GenreVI help me move all this mainframe stuff and translate it real quickly. Well, it doesn't necessarily do that better than the prior technologies we had to automate, , mainframe migration. But the interesting thing is GenreVI can really help you understand what those legacy systems were doing.
It can help understand the transaction logs, documentation, as well as the code and what was happening, and therefore help your engineers now Have an agent that can help them, forward engineer or modernize the app more effectively, which is a little different than people think. So it's Gen AI isn't the silver bullet that's solving the migration. It's helping our engineers in this human plus machine way now solve the problem more effectively.
So those are the kinds of things that I, that I'd suggest, , people think about as they look to build, , this business case to get to, , the modernized digital core of the future. fantastic there's so many people are gonna be, taking this excerpt and send it to their bosses or the board go listen to that if you ask me how fast i can go listen to that next time And to that point, , the question I'll add one more just in case people use it that way.
I think the question you should that every board or CEO or C suite should ask if they, they see a generative AI, , use case coming forward. Hey, we're going to use it to improve customer service. Can you use it for knowledge management? We're going to improve our supply chain over here. We're going to improve IT development with generative AI. The question I would ask, and I question every executive should ask is , not just , how am I, how are we achieving that? How is that?
Putting our digital core in a better position. Every step you take needs to move your digital core into a better position. I think that's a question that you should be asking. Otherwise you're just bolting on capability. That's going to make you more fragile in the long run. If every step isn't also improving your digital core, preparing you better for the future.
what are the things i just want to lean into and pull a thread on was the importance of every company is now a technology company because, many people will resist that. And I'd love you to explain what you mean, and maybe some examples of a company that saw itself not as a technology company, but now absolutely is, and you talked about the leapfroggers versus the laggards.
Yeah. You know, I started saying this back in 2012 when the title of our technology, we talked about technology vision earlier, the title of our technology vision in 2012 was , every business is a digital business. And as I went around on this, on the tour, talking to companies about that vision, I've, it was, I was widely attacked at the time because people were saying, , why are you saying that? My business isn't a digital business, blah, blah, blah.
Other than the industries, that were obvious, but within about 18 months. Everybody said, yeah, , that's right. Like, yeah, that's where it's going. And and that's where I started also saying if every business is a digital business, then every, every leader. needs to be a technology leader. Every CEO needs to be a technology CEO because technology, , the way you do digital, the way you use technology is going to differentiate your business.
And as a leader, you need to understand , how that's happening. I think that's become true in many organizations. Just a few examples. I'd highlight, , Walmart, maybe as one example globally of an organization who's embraced it going from, , going from , the, , the bricks and mortar store, stores of 10 years ago, it's an amazing company.
But, , look at what they're what they're doing now in e commerce and use of AI in their stores and use of data in very ambitious ways and kind of, , fighting, fighting against, , Amazon and the other other leaders there. Amazon themselves are amazing at technology, obviously, but it shows how Walmart's really transformed the CEO. Walmart, Doug McMillan has really embraced technology, really changed his team and changed the whole company approach to embrace technology.
I'd look at another example happening right now, the transformation right now, I'd say is in the insurance industry. Insurance as an industry I would say over my decades in tech has been a laggard in adopting technology. Generally speaking, insurance, I think most insurance execs would probably acknowledge that they've been a laggard in adopting new technology. It's borne out by, , the surveys.
If you look over the years, this time it's different, , insurance executives that the CEOs I talked to in the insurance industries are really looking at technology as a difference. How does this impact the way I look at risk? How can I use AI models in new ways? , insurance is largely a language based, Industry, it's policies and it's communication with customers and it's, it's it's, , regulation and stuff like that. And so it's really ripe for , , this new , development in technology.
So senior leaders are leading and they're learning about the technology. They're meeting directly with the AI companies and it's this change coming, , where that's cycling through the industry. So those are a couple of examples. And this is really true in every industry that the differentiation going forward is going to be how companies. Really , put this new digital core in place and use , AI, we're entering this AI era, use the technology to, to reimagine the way their companies work.
And that's what comes down to leadership. And we'll come back for those people interested in the likes of walmart there's a brilliant case study of arcado in the book as well and we'll talk about that when we come back to the ideas framework but i wanted to share by the way one of the things i do on the show i wear a pin to try and reflect the show in some way it's just one of these rituals i have to do. This is a kind of a human head with a, with data kind of a Rubik's cube stuck in his head.
So I thought it was like radically human. That's what I was trying to, trying to do there. It might've been stretching, stretching the metaphor, but, but radically human and human and machine has a, has a theme between us. And hopefully we'll have time to come to human and machine at the end and this missing middle. But in this book, radically human, you talk about innovation turned upside down and you talk about three major phases.
about intelligent technology and how we're entering this third phase, this, this radically human phase. But I'd love you to just give us a whistle stop tour of those three phases and what you meant by them. We started this in radically human, and then we built on it even more , in the new book, but the, , if you think about automation and people using technology, you can go back through Generate, you can go back industrial agriculture.
You can go way back with technologies, but look back to the industrial age. It was machine led automation. We're, it was assembly line standardized processes. Frederick Taylor scientific method. If you're familiar with these things, that was the machine led era that we started with. Then the second era came about with computerization, that I alluded to earlier roughly 70 years into that the computerization, , which gave us more, a little bit more, collaboration and processes and such.
And maybe you got to, you have human plus machine in some cases, still a lot machine led in terms of the technology, but , the technology fused together in different ways. And , in that era, what really mattered was we're moving from, standardized manufacturing, you know, types of processes , to human plus machine process that were put together, but there were still pretty static processes. Now, in this third, era that we're just moving into, it's really human
led . It's more adaptive processes. And that's really why generative AI is so exciting because it's technology that has many human like characteristics. It's got human like capability in certain areas. So rather than processes and interaction being defined by keyboards and different, , primitive ways of interaction. It's human level interaction with the technology that we're dealing with, which can really, give us superpowers in terms of, the power of people and the abilities of people.
And it leads to the need to change the way you think about processes. And that's again, why this word re imagination comes in there. And we talk about this this radical change , in the way we do things. A good example. of that, , human, lead thing, we talk about is a very simple example is in the design field. There's a a chair called the Elbow Chair, which won a lot of international design prizes several years ago.
It looks very different than any chair you've seen, but it's structurally amazing in integrity. Design is beautiful and it was designed by, , in a human led way. By a designer using generative AI, in this case from Autodesk. And the designer talks about this and says he would have never come up with the variations that that the generative AI came up with.
But it was then his human, it was his human prompting, then his human led, , filtering and exchange , and finalization that led to the award winning process. And that's what we mean by, kind of human led and more adaptive in the way it works. Paul, some people would be concerned about joblessness, even in this age of human and radically human technology. And if you think about the jumps in even the GPT releases for those people who use that, there's lots more AIs out there, but.
mainly people have adopted Chat GPT i think but the jumps in each of the iterations are huge and when you think when you roll out forward to another ten years ago where is it gonna be and. Is it a case of we're training the AI at the expense of ourselves? Are we digging our own graves in some way? And you mentioned Taylorism.
I often think about technological Taylorism , like the way Taylor used to sit and look and see how long it would take you to put a piece together, whatever that AI is now watching people and seeing. How quickly they do it or how they do it or else they're even teaching the machine you talk about machine teaching for those people who are concerned about that what would be your feelings.
This is a long topic, but the What is your happy to get reason, it's the reason we wrote the first book , , even before General AI came along is because there was in the mid, in when, with the advances in deep learning in the early 2000 teens, 2013, 2014, same discussion was happening. And in fact, Jeffrey Hinton, who recently won the Nobel Prize, , Jeff Hinton famously said in that era, we need to stop training radiologists today, because , in a few years, we will need none of them.
Brilliant guy, won the Nobel Prize, totally wrong on that topic. I think a lot of people misinterpret that. The way that this technology will really be applied in the way that people work. And and that's why we wrote human plus machine, because , we think , we need a roadmap to think about how to apply this to achieve the right impact. We also did some research and we're publishing this in some , new articles coming out in Harvard Business Review.
And there's been some other studies from MIT and Stanford and other institutions about why human plus machine leads to better results. And One of the examples we talk about is in a chemistry type of process, a researcher on their own had about, 60 ish percent , accuracy and ability to get something done. AI model in this case, using something like AlphaFold , from DeepMind from Google was in about the same range. Little bit better, but about the same range together.
It was like 90 plus percent accuracy in what they could do and that's been repeated in many, many fields. And it's subject of a great paper from M. I. T. coming out now. And that really is the road map. We're trying to provide a human plus machine. If your goal. is to use generative AI and try to automate what your people currently do. Go for it. I'm sure you can do a lot of that, but is that going to position your business better?
No, because the real, I believe the real differentiator will continue to be how your people use the technology and driving your business. And that's what organizations need to focus on. I, another thing I've been saying for a long time, For, , 10 ish years is, the companies that will be successful are those that invest more in the people than in the technology, as much as you need to invest in the technology. So. , That would be the philosophy.
Now that said, there's some big problems that we have as a society and as organizations to face, I believe there'll be more, there'll be more than enough jobs and just look at, what the internet was supposed to do in terms of displacing jobs and look at the number of new intrapreneurs on platforms like GoDaddy and eBay and, Other platforms around the world. Millions, millions, millions of new jobs and businesses being created in new ways because people are empowered with new skills.
That's going to be multiplied exponentially with the human capability of generative AI. So the only way you have to worry about the net total of jobs is if you believe every job has already been created. It's a zero sum game. We're just. Automating the jobs of today, and I don't believe that's far from the case. I think we probably only invented 5 or 10 percent of the jobs, the jobs and services and products that we as humanity need.
So that's the philosophical , type of answer type of answer to it. But the practical answer is then to like work on the human, the missing middle and the human plus machine jobs that will create the opportunity of the future. The other thing I'd say though, a big issue is we already have, and this isn't being talked about enough. We already have too many people on the wrong side of a digital divide.
Whether you look at global South versus developed economies, or you look in individual countries, socio economically or. By different groups from an inclusion, diversity perspective, we already have a digital divide that's unacceptable and leading to political issues and unrest and various things that gender value and all this technology will make that worse if we don't do more about it. Those on the wrong side of the digital divide to start are going to be.
Further on the wrong side, and that is a big problem. I don't think that's being talked about enough. We're so distracted by artificial general intelligence and all these other things that I don't think are the real issues, that we're not spending enough time on the basics of. Let's help those on the wrong side, digital divide, get the right skills so that they can catch up and be effective going forward.
It's one of the reasons why with all three of the books we're donating, all net proceeds, all royalties. to non profits that are helping mid career displaced workers, because we believe those are the ones that, that need the help most. And we've been able to support many great organizations great organizations with that along the way.
And i just wanna say as well well now i kinda feel bad saying this cuz i want you to buy the book but i have a copy of each up for grabs if you just follow the innovation show Substack i have two copies of for grabs one of radically human and the other of human Just, just buy ten more, because it all goes to a good cause. You just heard, seriously, we've been able Yeah, man. I, so I, great, organizations. well done.
Well done. That's Bravo because I'm sure you've sold a hell of a lot of these as well, particularly you have a second edition and you've more on the way as well. , I know you've way more in you. So. I thought we'd move on to the framework. I, by the way, I hope you, I hope this is right. I was, I spoke on the show to Yossi Sheffy, who's one of the world's experts on supply chains. And he had the same feelings. He was like, and Yossi's still going. He's amazing man, still lecturing.
And he said the same, he said, Aidan, your fault. Cause I, I'm a little bit like, Oh, I'm not sure about this , is a business, a platform business. If it's a platform business, wall street rewards it for having higher. Revenue per employee. And I'm like, Oh, this is the formula is wrong here. And particularly with the digital divide and he's like, Oh yeah, but you're.
You're making the assumption by looking at the past and that's what everybody does in each paradigm shift is they look to the past to make decisions going forward and you're falling for it here so i hope that's the case and i really do and i hope those people who maybe are like me or inclined to think this is a dangerous time, maybe i'm mixing it up like you say with the political divide and the digital divide with people as well because you see that quite a lot so let's hope it's the case.
Well, just one more thing on that. You just triggered another thought with what you said. And yeah, he's an amazing guy. So it's good. Good that you you talked to him to the what we're seeing. We just did another piece of research. I'd be happy to. I'll send it to you. And you happy to free to post it to your listeners. The title of it is work, workforce and workers is the title of it.
And the point we make in this research is that we have a bit of a trust gap forming in the workforce, which is another issue related to this not in this serve in this research, you'll, that you'll see in the report 94 percent of workers, this is across all categories of different types of work across the spectrum, globally, 94 percent of people believe generative AI is going to help them and be good for their career. And that's going to benefit them.
94% about the same percent believe they can learn it. It's not going to be an issue to learn it. But when we talk to leaders, it's almost the opposite view. Leaders are concerned their employees can't make it. And uh, they're concerned whether they have the right workforce and that's building this trust gap, the workers believe it's good. The leaders aren't sure. And when we asked the leaders, you know what they're doing, they said that they don't know the impact yet.
So they're not saying anything. So there's a silence coming from leadership workers believe that can do it. And as a result, there's this distrust building. So if you ask the workers how they're feeling about it, they say, well, I'm optimistic. I don't believe my current company is. Yeah, I believe I'm at risk in my current company because of it, which is just this great paradox and this great lost opportunity to maximize human capital and why this human plus machine thing is so important.
And those organizations, those leaders who are out there talking about you might not have the perfect answer, but talk about what it's going to have. Talk to the employees, educate all of your employees in What this is and what that's doing, encourage them to use it so that they're part of the process.
Those are the organizations that are reversing the trust gap and instead building, a workforce of employees who have the trust and also will bring the ideas and the culture and everything else you need to drive the change. Those who are stuck in this, trust gap thing, I think we're gonna have a lot of trouble going forward. So there's, again, advice to organizations. And then it's, what do you do about it rather than talking about it?
You need to invest in your employee skills, build durable learning platforms because , people are gonna need to learn and relearn and relearn. And less than 15 percent of organizations , are building those kinds of learning platforms through employees now. So that this is, this all needs to change as well. So maybe I'm getting back to the, While you're pessimistic a little bit, but there's also the action that you need to take. You need to have a vision and some message on where this is going.
You need to talk to your employees, you need to invest in the skills foundation. That's doing, which gets back to this idea that I said earlier, you need to invest more in the people than the technology. I think that's such a massive gap. And I get it. I mean, organizations were in this real paradigm shift moment in the business world, and then you have younger generations coming through older generations will maybe argue sometimes out there.
They're very demanding these younger guys cause they want progress and they want it learning and they want budget to be able to learn and they don't get it often times. One of the biggest fascinations, Paul, of me was over 10 years ago. I left a role and I was so frustrated because I couldn't get funding to do a project. I wanted to do the prince to project management.
Let's see if i wouldn't let me do it and then he gave it to me for buying a book on the company credit card i was like boy am i in the wrong place cuz.
It was like i just saw that you have to have a certain level of learning in order to be able to evolve in an organization to be able to evolve your thinking and if you're not investing in people as the world changes their left with these legacy, this legacy junk ware that's no longer fit for the world and i think it's such an important point about this constant investment in your people so they can keep up with the changes in the world Speaking of which one of the ways your
people can get a grasp on all these things changing the world, is the ideas framework and i love the ideas framework because it's the really can be the building blocks of how to navigate this change and it's really simply, Evolved it's really simply built in order for people to be able to get it so you're not trying to complicate this at all i'd love you to firstly at a high level take us through, just what the framework is and then i'll pull on some threads and we'll go a little
bit deeper into each of the five steps I have to give credit to my coauthor, Jim, for coming up with ideas. We had these five things and we're trying to figure out how to make it memorable for people. And he was the one who came up , with the acronym ideas. . So the five things, so it's ideas, I D E A S, so. The I is is about intelligence and and we use a tagline of, , more human, less artificial.
And this is the point we were just talking about of, we're making two points there, by the way, radically human was written before gender of AI, but we were talking about the shift. One thing we're making the point of with the eye here, more, more human, less artificial was that the technology was becoming more human. We were kind of looking at transformer and these things that became.
The T in shot GPT before before that moment, and so we're kind of anticipating the more human like nature of the technology and therefore why you needed to embrace these more human ways of using of using the tech, the technology. So that's the, The D part is data and we're making the point from maximum to minimum and back is what we say in the book. At the time when we wrote the book, everybody was just focused on the big data.
And we were seeing things like zero shot learning and other techniques, which were about how do you create a path when you have less data? And there's other optimization techniques and such that are very good at less data. And then there's synthetic data that we talked a lot about in this chapter. What do you do when you don't have the right data? What if you need to train your manufacturing system to identify defects, but you're so good, you don't have many defects.
You don't have enough data to train it on. Well, you need synthetic data, to train it. And so we're talking about, don't just get enamored with the bigger, the bigger, the bigger. Think about the best data and the best way to use it, the best techniques, because as we know, Using a lot of data to train these models costs, , millions, hundreds of millions of dollars. Tremendous energy impacts and all the rest and using, finding the right techniques for the problem is really important.
So that's the point we're making around the data. Expertise is is talking about moving from machine learning to machine teaching. It's just a clever way of saying try to reverse the equation a little bit. And a lot of people are fond of saying, well, you need to keep the human in the loop. With it with AI. I just think that's backwards. It's not like this tech overlord and the human needs to be a minor part in the loop. It's the human is in control of a process that includes technology.
That's the way to think about. That's the point we're making in the expertise, . One of the points we make in that section, Machine Learning to Machine Teaching, is what's the role of the human in this, and how do people make the technology better? A great example of this right now is what we see in life sciences, when we see AI being applied by scientists. Scientists have a hypothesis, learning models, starting, , starting the models down a certain path.
The model generates some insights, the researcher says, aha, yep, That's interesting, but let's go a different direction and the model's learning and the scientist is learning and they're coming up with better results than either could do on their own it, as I said in the data I cited earlier. Architecture is talking about going from legacy to living systems. And that's this digital core idea. How do you move from legacy?
To living and by living, we mean dynamically adaptable systems with the properties I talked about earlier. The modern data application and architectural foundation. So we can talk about this a lot more as well if we want to dive into detail and then the strategies this is kind of a fun part of the book to work on is, and this came from interviewing a lot of leaders and organizations is is yeah.
If we're, if every company is a technology company, then how does your strategy need to change to incorporate AI and everything that's happening around technology? And that's where we talked about some new strategic concepts Forever Beta, Minimum Viable Ideas Colab, a few different frameworks of strategies that we see companies that are very successful adopting. i love them and i love those ideas the minimal viable idea colabbing, hopefully we'll get there man hopefully i can fit this in.
Thank you for giving an overview of ideas. I noticed sometimes looking at the data to show that people kind of get distracted and might drop off halfway through. So I might cut it even into two pieces. So people get this right at the start, the ideas framework, because it's so, so valuable. The intelligence part, I just, there was a couple of things here that I don't think most people knew, which me included by the way, that we hear a lot about deep learning.
And firstly, many of us don't know what deep learning is, but certainly, we don't know that it's kind of has limitations. And this was one of the big lessons I had from the book is the limitations of deep learning. I'd love you to unpack that for us because that was a real eye opener. Yeah, the, I mean, so deep learning it really took off in the two 2010s.
Jeff Hinton, who I mentioned earlier, that's, you know, part of the reason he won the Nobel Prize, was he was an amazing pioneering researcher with some new a lot of the new concepts around deep learning. But, deep learning does, there's lots of different forms of deep learning and yeah, you do, supervised learning, unsupervised learning, reinforcement learning, reinforcement, human feedback, all sorts of different techniques that can help you improve the results.
And yeah, you have to really understand, the right techniques to use. And, at the same time, it's, it's sometimes not, the right, approach to use and it's got, , limitations in terms of the amount of data and in such that you need to use it. The emergence of generative AI, gave it new life, so to speak, because generative AI and the models are built with some of the techniques that I just talked about, but with the innovation, the added innovation.
Of transformer technology, which is a new way, kind of a creating and predicting what might happen next, which is what the models do well and diffusion models, which are used in the multimodal that more visual models that those two. Innovations added back into everything that was happening around deep learning is what led to the, you know, the boom in generative AI and kind of, you know, fueled where we are today.
But one, one thing we talk about is, is there's things if you're trying to predict your use AI to predict your retail sales in a certain market. 18 months from now incorporating economic patterns and weather in your business trends. Deep learning is probably not the right answer. There's other forms of machine learning and optimization problems and such that are, have been used for that for a long time and are appropriate to use.
If you're doing a next best, next best action kind of things, you know, if you're on Amazon and trying to figure out what book you're going to read next, that's probably not going to be deep learning. It's probably going to be other, other forms of, of machine learning and and AI that are, that are used in there. So. That's one of the things we're talking about in the book is making sure you're choosing the fit for purpose types of models.
And we talked about really three, three broad categories of AI, the diagnostic AI, often deep learning driven predictive AI, often driven by optimization and other algorithms, and then generative AI with a unique, now creative capability. You know, powered by these, these new types of models.
One of the things you mentioned about deep learning is that , it has a black box problem and that it's making really consequential decisions, but when you try to figure out how to make those decisions, it's pretty opaque. That was one thing. And the other thing was deep learning systems can't read. So when people are actually uploading documents I think this was really important. It can't read those documents. And I'd love you to share a little bit about those two things.
Yeah, that's where there's a lot of exciting things happening now too. So yeah, deep learning kind of on its own, you have , this black box and explainability problem. And there's ways to solve for that. There's general Adversarial networks and things like that you can use to try to understand what they're doing, but still approximating some of the interesting innovation going on now is how do you take, how do you add the explanative explainability back in that?
There's some really interesting things happening. If you use perplexity, for example, one of the generative, big generative AI models, you'll hear headlines as this comes out about the large amounts of money they're raising. Perplexity is one of the successful. Frontier foundation models. They have a explainability built in when they tell you something, they'll give you the citation is this came from X, that, that's really important to do.
And others are working on that type of problem, but deep learning on its own doesn't have that capabilities, to, to your point which is important. So , that's and one, one thing I've, I was just working with a company on is using that type of model in an investment. Portfolio analysis application where you don't want to, you don't want to have hallucination make you make the wrong bet. You want to understand as financial analysts, I guess, human plus machine.
I want the model to advise me, but I want to understand and make sure, make sure it's the right bet. And that's why the explainability features come back in. , that's really important. Working with this company on on recently. And the second part, so it was explained, but it was the second part was the, that I can't read, yeah, and that was a big limitation. And, you know, that's why you have to look at how you combine different techniques.
So you can use, now you can use NLP and generative AI, you can use them to do the reading and adjust, and then combine it , with deep learning models. One of the, it, and But you have other problems. The generative AI, the new foundation models, aren't good at math, so they're because they're kind of language based.
So you really have to understand, , there's no one silver bullet for everything and the answer, we can come back to this, but the answer Isn't just making these models bigger and bigger with more hundreds of billions or trillions of parameters, that's not going to solve it. It's kind of combining these techniques and models in different ways. So for example we we did a man, a manufacturing inspection system for one of the largest automotive companies in the world, almost more than 10 years ago.
And it used visual use cameras, video cameras on their assembly lines with a with a deep learning algorithm to identify defects, back, so this is with The common technology now back then it was more pioneering, but now with generative AI, we can do some new things. We can understand the nature of that defect. We can look at documentation and other things.
So we can use language based resources now, understand, and add to, the power of the system that can generate the defect and kind of look at all sorts of other dimensions of it. So it's sometimes looking at how you combine You know, multiple dimensions of the, of AI together to solve one problem. It's not like generative AI displaces the prior forms of AI machine, simple machine learning optimization algorithms are relatively simple in many cases are going to be the best answer.
And maybe you can wrap generative AI around for better human, interaction and understand the context of the problem and such. There's so many great examples in that i'm actually look at my notes here going i don't have enough time to cover all these examples i love the stuff the covariance in there i tracking by etc etc so i leave that to people now and now particularly that you know where the proceeds of the book go to buy the book so we better keep moving on to data and.
Again, lots, lots of huge insights reading this chapter, particularly things like the lack of big existing, big data sets for AI. So maybe we'll start off there and maybe you'll give an overview of the world of AI, of data. And then I'll pull on a few threads. You know, when you look at data, most companies understand and can itemize their transactional data. So, my customers their history, billings and such, and that's one level, but even that in many organizations has issues.
You have, no single source of the truth and things that, that need a lot of work, quality of the data, et cetera. Then you have this, the kind of second. Type of information you're dealing with, which is the unstructured data, your call transcripts the what information externally, what people are saying about your company and on, Yelp or customer view sites or what have you.
So unstructured data that may come in different ways, and then you have synthetic data that you may be able to create that will help your business in different ways. That's data. And then the other thing that you need in that you can use that in as a foundation integrated to power the AI that you need, but then sometimes you also need data sets to drive the data. Think about scientific discovery.
You may need , things like like alpha fold, which has the protein, folding predictions and information and such, or if you're looking at one thing we did for the book is we have some AI looking at Job classification and job trends and what's happening and use of technology. Well, there's massive data sets on that fortunately available that are aggregating a lot of different company information, but that data doesn't always exist.
And I think the lack of data sets is one thing that we're seeing organizations deal with. And in some cases, we're seeing, you Industries collaborate. Leaders in industry say, Hey, let's put this data together.
You know, clinical trials around related drugs in the pharmaceutical industry, for example, if it can help all of them, you know, with different aspects of the problem and, you know, avoid some of the, you know, the intensely competitive areas, then there's value in collaborating to create data sets that will help them power. New things going forward.
So that's, that's, you know, there's a lot of work going on on this because, you know, the, the, the AI will only be as good as the data in the data sets that used to train it and they feed it to produce the results. again, with some empathy for your colleagues in CTOs all over the world who are pulling out of clips and send them to their bosses. One is. one thing that is that that's a new thing to look at there.
And some on the call probably are looking at this is how to use AI to improve your data. There was a great study, not to say Walmart again, but just just happened that one of the interesting that, that. Interesting things they're doing with AI that they talked about recently is how they dramatically improve the quality of their product categorization data that drives all their online business.
So in other words, this is, this is the data that's really important to make sure if you search, you find the right product that you're going to buy. They have some quality issues with the data and they use AI to dramatically improve it very quickly. And it's an example of , of of using the technology to improve your own data.
What are the things i was gonna just mention was the rising cost of data so training a single model is relatively inexpensive but the better you get the more it costs and i thought that was important thing to mention, Yeah, I think what's one of the issues that's going to drive how AI evolves in companies. We're at the very early stages with these large models.
And and again, it costs hundreds of millions of dollars to train the biggest models and therefore, the commercialization of that to use those models is going to, which is called the inventory. There's the training cost and the inference cost. A lot of times those training costs hundreds of millions of dollars absorbed by the training. The frontier model of companies.
But to use those models, it's the inference costs that you as a customer pay, those costs are very expensive for the largest models are measured in the tokens you submit or the input you give and things like that. So, it's a big consideration for companies because yes, you may be able to use a very sophisticated model to get a good answer to support your customer service process, but can you afford to do it? And that's And that's an important consideration for companies going forward.
Similarly, if you're developing or tuning a model for your own use, how much is that training, of the model for your own use versus the inference, versus your use of it. So these are, we're at the early stages. Companies are starting to understand this. We've developed frameworks to help, companies evaluate , these types of of costs, but that's why it's very important.
to make choices, only use the really powerful models or only build those really powerful models if you really need them because they're incredibly complex and expensive to use. Where can you use, an open source model, say from a meta, one of the Lama models or Mistral or other organizations who have much smaller, more fit for purpose models that might help you in a domain in a more efficient way than, the bigger models, but that's, it's all sorting itself out right now.
And that's going to be a really important area to stay on top of over the next few years. One of the things i thought was really important was again you mentioned here, even leaving budget. We're in budget season for many companies and they don't leave enough budget. They think they might have enough budget to do these things. And then they have this problem with data. And you tell us that AI pioneers are addressing three distinct challenges. One is to find the needle in the haystack.
You've a bunch of data, but what's relevant. The second is the amount of data. Is small so we talked about this in the intro the efficiency of that data and then that there's some cases there's no data at all synthetic data for example you mentioned i'd love you just to give a quick word on that, One of the things that you need to really think through is how to integrate the data in the right way and how to use some new ideas.
So there's vector databases and knowledge graphs in different technology. Some of you will be familiar with that. The rest of you will become familiar relatively soon with it, I'm sure. The and it's, these are ways to take the data you have, which in some cases may be small data. It may be.
Specific supply chain data about your business in the way you build products and such, which isn't, , massive big data, but it's really important to the problems you're going to want to solve with AI creating, the knowledge graph and, vector databases, vectorizing the information you have. So that's really important. Easily available and digestible by the larger models.
These are really important areas that are going to, that are going to dictate how successful you can be in getting accurate results and getting results that are really tuned to your business. So thinking about your data, as I talked about earlier, all the different forms of data, and then thinking about how do you prepare that in the right way to interact with your gender of AI and the other forms of AI is part of the art of getting that architecture right. We better keep moving, man.
Cause I'm going to run out of time here and people are going to be left hanging. We're on E, which is the expertise piece.
And this is the move towards machine teaching human and machine will cover that again in the future but in this book and human and machine you talk about the human machine hybrid activities, that you've identified and maybe we'll just give a quick overview of your favorites of those Yeah, I think, when you think about the expertise, there's really You know, from machine learning to machine teaching, there's some simple, ways you can see this in terms of people labeling and tagging things.
That's really not what we're talking about there. Those aren't necessarily the really high value examples.
The other types of examples that you see here are in areas like customer service, as an example, where think about a large telecommunications company we worked with where They use, they have, customer service agents, but they also have AI agents that can listen to the call, understand what the customer might be asking about, and then, suggest what you have to the human agent, what might help resolve the problem. In this particular case.
The use of AI and generative AI agents increased productivity by over 30%, increased customer satisfaction by over 60 percent because people are getting, the customers are getting better answers and it had this amplify effect. If you think about the, what we were just talking about Aiden, where. The AI made the person, you made the human agent better.
They were able to, that was, as far as the customer knew, they were just dealing with a human and that human was more effective and had these, superpowers to answer the questions better. But the AI was also learning based , on the interaction and what the human did. Did the action resolve it or not? So that next time around, it'd be that much better. And that's that amplify and interact, capability of the machine teaching coming from the human interaction with the AI.
Paul mentioned in that brief overview earlier on architecture. And what I found really interesting was that architecture in your IT system is starting to reflect art architecture in the organization Conway's law, where. We need to be more fluid we need to be more adaptable. And I love the way you talked about, for example, the top people. So the AI pioneers are not only using the cloud, but they're innovating in the cloud.
And I thought that was so interesting to show people , how far you can go to be an a student. Yeah, that really is I think for the innovators. And again, we talked earlier about the, , the way that they're outperforming. Part of that part of their whole rationale for getting to the cloud was it was the only place, to really innovate. If you're not in the cloud And there's certain things you can't do in the cloud for a variety of reasons. I totally get that.
But will continue to see more and more move to the cloud. About 40 percent of enterprise workloads are in the cloud. I think that'll double over the next few years, , is what we're going to see. And a big part of that is innovation. If you're not in the cloud, you're in your own environment. You're trying to innovate and add new services and features and keep up so that your applications can, can take advantage of these features.
Why not be in the cloud where, the cloud providers are introducing these new things on a continuous basis. And that's the insight that some of the early leaders that went into the cloud had and why they're innovating really rapidly. So if there's a new, AI capability, so if, you happen to be a Microsoft customer and they release a new, open AI capability in Azure.
You can take advantage of it right away rather than needing to figure out how to deploy that in your organization, which therefore focuses your teams more on the business innovation you need rather than, doing the nuts and bolts of the technology work. And I think that's a core insight that the leaders have had and that will continue to drive it. As we go, , now we have cloud, we're moving to the edge, we have AI layered into it with all these new technologies coming about, the more you can.
Leverage, the partners who provide that technology rather than building it yourself, the more you can invest the precious dollars you have on the innovation that's going to benefit your business. So we're gonna have to move on and i'm gonna use the architecture and use actually everything i'm gonna use idea to finish on the s which is strategy, what my favorite chapters in the book brings everything together great case study you given that is the lemonade case study i'd love you to share that.
Yeah, it's a great example because it's, it's also showing, how it, , how you can just approach an industry entirely differently. So Lemonade is, for those of you who don't know, it's it's really a startup insurance carrier who had a different proposition about that about how to approach the market. It really sh it's really. Minimum viable ideas framework, which is taking like the different elements of ideas using the intelligence, the the data, the expertise, et cetera.
And so what Lemonade did is they created a, cloud native new business powered by intelligence, you know, getting insights quickly, online immediate response, to policy requests and things that, that the customers were asking for. And and they had, humans in the loop, humans as part of the process, , powering all that.
And and they've been very, they have been successful with this whole new model in the industry, kind of re imagining what insurance might be for, , a different segment of customers who wanted this very fluid, digital, adaptive process. And I think it's a great example of a whole different approach to strategy and that's, and that they continually iterate on and improve on as they grow their business and expand it.
We have to cover man i'm gonna squeeze it in i love minimal viable idea forever beta and colabbing as well. Got to share them and was squeeze them right at the end where people's appetite So forever beta is another, another one of the strategies that I think is, is really resonating that we see, companies adapting the example we talk about in the book a lot, I think it's just a great example is, is Tesla who really that their car itself is a beta. I drive a Tesla.
, some of you in the audience may drive a Tesla. And I got a recently got a new download of a vastly improved driver assist, near autonomous driving capability that they downloaded. So they view that the car is the platform. And it's, , it's forever beta, forever improving the product based on customer need, based on the improvements of the technology and such.
And I think that's the best example of companies that live , that that model are going to be able to, to continuously adapt and delight their customers with these new experiences.
Another one that follows that, I think it just, It's a single example would be Amazon, , but the Amazon Web Services in particular with it again, with Jeff Bezos famously had, that's, that's always day one kind of mentality and that led them to this to this mindset where they continually introduce new products, new features, new services to delight their customers. And then, so that's, minimum viable ideas. And then Colab is the other strategy, which is when you really.
Focus on the collaboration, the co lab and the co lab had a science connotation. The way we talked about the book too, as well as collaboration. And I talked about examples of this a little bit earlier. We see this in particular when we see this applied in areas with scientific, , research, , types of areas where we see this happening. Moderna might be a great example of the way that they pioneered the COVID vaccine by pairing their scientific processes with AI powered models.
They had an AI workbench that they used their drug discovery process that allowed their scientists to innovate through all the permutations of the virus and potential , new new variants. RNA based therapeutics to come up with the right answer much, much, much faster, , than anybody else could have. Could I just absolutely fantastic want to tell the audience we only got through the first half of the book because in part two of the book.
Paul talks about how companies use the ideas framework to differentiate themselves on four dimensions talent trust experience and sustainability and it's absolutely brilliant it brings everything to life and we have to finish it there and i hope to have paul on again to cover. His other book, human and Machine, it's been updated recently as well.
just add one point just from that back half, just one point I'd want to emphasize is the trust point since it's come up and we've hit on it a little bit as we went through, we talked about a trust gap with employees and such, but I would, I'll just make the point , for, Listeners and viewers that trust is one of the key things. I'd encourage you to look at that part of the book.
It's in, there's a part of segment on trust and all of them, because I think trust becomes the differentiator in this next era we're moving into.
If this is all about data, and if this is all about creating new experiences for customers, and if it's about more human like capability, those companies that really instill trust and get trust within their workforce, trust within customers and trust within , their stake, their stakeholders are going to be the ones who have the right and the currency to move even faster. Those who violate the trust, customers are going to opt out.
Employees aren't going to want to, work in those types of environments. And so trust, I would strongly suggest is an area for everyone to really think about. And how does every part of what you do in ideas lead to trust among each one of those groups that you're working with? Beautiful Paul. Nice way to finish, man as well. For last question for you, for people who wanna reach out to you or find you, you do keynotes all over the world as well. Where's the best place to find you? Do?
Best way I'm on LinkedIn, active on LinkedIn. Reach out to me at Paul R. Daugherty. I think you'll you'll find me out there without too much trouble and love to hear from you. Author of human and machine. Hopefully in the future, we'll cover that. And this beauty, radically human, Paul Daugherty. Thank you for joining us. Thanks, Aidan. It's been a pleasure.