Alexander Sukharevsky, Senior Partner and Global Leader, QuantumBlack, AI, McKinsey & Company
In conversation with: Alan Murray, FORTUNE

Fortune Global Forum 2023: Generative AI - The Next Productivity Frontier 

Conferences

Alexander Sukharevsky, Senior Partner and Global Leader, QuantumBlack, AI, McKinsey & Company In conversation with: Alan Murray, FORTUNE

Fortune Global Forum 2023: Generative Ai - The Next Productivity Frontier 

Now let's talk about another elephant, yet another elephant in the room. Lucky. It's a big room here which is rapid technological advancements they have of course, come to impact, inspire and yes, alarm people all over the world. Probably none more. You could argue than generative A I I think we've all seen that particularly in the last week with open A I. It is a powerful tool, of course, uh becoming more and more ubiquitous every day. According to mckinsey, our knowledge partner, generative A I get this could enable automation of up to 70% of business activities, 70% across almost all occupations. So whatever you're doing, watch out between now and yes, you got it 2030 which could add trillions of dollars to the global economy. But of course, also displace a lot of jobs and the tech industry says that this is only the beginning. So how and when should business leaders take steps to get into the action and implement A I, I'd say the best time was probably yesterday and the next best time as they say is today. But we can get some insights on this question from Alexander Sukurs who is a senior partner and a global leader of Quantum Black A I by mckinsey and company and our own Ceo Alan Murray. Thank you Alexander. Thank you for doing this. I I think generative A I is sort of the topic of the year. I have lots and lots of questions for you. I appreciate you taking the time to do it. Thanks for having me here. But before we dive in though, can you explain to people what mckenzie quantum black is? That's what you run? Sure. So quantum black is one of the largest A I companies on the planet with 5000 colleagues in more than 50 countries. Five are in these centers across five countries as well as an ecosystem of external partners where, where we have the IKAR is clearly Stanford University for Human Artificial Intelligence, but also the upcoming leaders in generative A I space. So it's a big player in this area. Yes, one of the largest ones. Yeah. So, so look, I I had never heard the term generative A I until about a year ago. So this has happened very fast. I've spent four decades watching business adoption of technology. I don't think there's been a new technology that exploded in business as quickly as generative A I has. And I think the question everybody's asking at this point is, is this real or is it hype? I mean, what's the, what, what's your sense from all the companies you've worked with. Well, I think, first of all, it's real, but we are extremely early in our journey while we all believe that it has been a while. And actually what the concept that was introduced in 2017, essentially only in July beyond a few privileged organizations, 10 million software engineers got access to these tools. So whatever we are seeing today is actually pre awareness and pre deployment. On the other hand, and you're right, we got it wrong many times with dot coms and crypto and few other things. But I truly believe it's the first time in the last 15 years that we see a real platform shift as platform shift happens when the cost of production goes down 1000 times. And if you were a comparison similar to what internet did to distribution, generative A I does to creativity, meaning the marginal cost of producing a movie, a slide, a legal recommendation goes close to zero in significantly faster rate. That's number one, number two, we are actually able to interact with technology in our human language, right? So think about you and I having our virtual assistants that schedule our day shop for us, provide us emotional support when we need. And also in the organizational context, we manage quite a few virtual assistants to help us to do our work. So the combination of really different cost curve and being able to interact with technology in the human language, that's what makes it? So that's the potential, the potential is huge. But how far along the curve are we towards reaching that potential? Right. And I think you are spot on here, we are just at the beginning and your colleague mentioned the number of 70%. So if we look around this audience, 70% of employees task today could be automated in 20 years, 50% of them will be automated. At the same time when generative A I appeared last year, the magic number was 11% only 11% of use cases ever made it into production. So one out of 10 is real. Now, if you fast forward to generative A I, considering some risk considerations and cost consideration, this is a single digit number. So what we are trying to do within quantum black is really bridge this gap between 70% promise and single digit implementation in the enterprise context. Yeah, so give some examples of where you see the big opportunities actually going into production. So I would say, although 90% of the world is focused on productivity and I would touch upon it. I do think the bigger thing is actually to step back and considering the changes that we just discussed to think what is the mode for your business? Because all of a sudden you allow to intermediate, this intermediate businesses, you are allowed to deliver the same value for few cents the dollar and you're able to acquire new users pretty much for free because if you look at recent generative A is their RP is three times higher and the traffic is organic. So I think while we all focus on productivity and playing with different tools, it's really trying to reimagine your business and industry. That's hard work. I mean, it's easy for everyone to talk about how they can improve the productivity of their software development of their marketing department. But you're saying you really have to re imagine the business. Exactly, to avoid disruption. Exactly. Because while the new kid is generative A I it's only 1/4 of potential of A I. On the one hand, on the other hand, you have seen quite a few other technologies that get mature. Now be it virtual reality, be it web 3.0 round toch organization. So if you were to combine these three and think about what does it mean for your industry, it might be a consumer industry, it might be a basic material industry. I think the shift is quite fundamental. How many companies do you see actually doing that? Fundamental rethinking of their business? So, as I said, unfortunately, it's 1 to 10, probably 10% of the companies are focused on fundamental rethinking while others might arrive there, but start with the productivity. Yeah, so we still I talk to ceos about this a lot and, and I hear a lot of caution concerns. They talk about data protection they talk about IP. And then the one that really scares me is the continued hallucination, the tendency of this technology to make stuff up. It's very scary. Like what, how do you, how do you deal with those very real problems? Right. So I think I'm less fond of the notion of artificial intelligence. I'm more into hybrid intelligence, bringing the best of humans and machines. Now somebody says that hallucination is a feature and when human being hallucinates, sometimes we don't pay attention and we actually ignore it. So I think while we very focused on a few of the risks that you describe in the short term, I think it's important to have human in the loop. But I do believe there are quite a few powerful tools that could allow you to protect your IP protect the data, ensure that the machine works in the right way. Again, having human in the loop. What are the things that I'm more concerned as a leader in this sector? Also as a human being is the longer term implications and some of them are unit, some of them are esg implications. Some of them are ethical. I think we as humanity need to decide what we would like to take and what so highlight the two or three top concerns you have about the long term term implications of this technology. No, I think first of all, it's the ecological use of the technology and application for environment. Second thing if by that you mean to mitigate environmental problems or are you talking about the compute use to run these large models? I'm talking about the balance of the contribution. Basically, how do you minimize the harm to the environment? On the other hand, you solve the most difficult monar problem that actually we as human beings cannot solve. Right? Second thing is how do you really empower human beings with this technology? How do you allow the superpowers that discuss your employees to develop themselves and just to become stronger with the technology? Yeah, other impacts on society that you're focused on. Well, I think as I said, the moment each one of us is able to interact with technology in a human language. And all of a sudden we find ourselves working not just with human beings but with virtual assistant. I think both the corporate rules or the governance rule. Be it the societal rules be it? Let's think about retail. So think about my virtual assistant shops with your virtual assistant and they're trying to find what exactly I need and provide me these goods. What are the implications for advertising industry? What are the implications for bricks and mortar? What are the for my own choice? And how do I train my virtual assistant to actually act in my best needs? Yeah. So what are the industries where you see this happening fastest? You just talk about retail but give some other examples where you see big implications coming quickly. I think it's quite fascinating because if you look at the impact that we discuss, it's around four trillion that is essentially bigger than GDP, all the United Kingdom. It's massive. And I think it's what is one trillion over what time period? If we think about the theoretical potential that could be actually achieved today, if you kind of remove all the barriers. Now, what I do believe that if you look at our research within 20 years, 50% of the potential will be realized. Now, I think what is important that essentially if we and again, we are now focusing on productivity, not reinvention of the business because that's where you could put a number. Uh There are four use cases that provide this potential. One is concision ability to take any database to summarize it and to interact with it. So be it blue collar workers for a mining company, be it knowledge workers, it works. Second thing, it's really knowledge creation and content creation where you could create a picture, sound movie essentially with a single scriptwriter number three. And I think that's we also is really coding and software development. They take legacy into modern software or accelerating development. And then last but not least, it's all that has to do with the customer interface and customer journeys. So four C is essentially that brings most of the value. Yeah. So let me ask you a debate that I have heard uh conducted between CEO S. There's one group of CEO S, let's say more legacy companies who look at this and say, hey, this is really interesting technology, but it's also a little scary. I'm worried about my data. I'm worried about IP I'm worried about hallucinations. I think I'll be fine as a fast follower. Let somebody else go first and hit all the landmines and then I'll go And then there's another group who say you can't do that. If you're a fast follower, you're going to be a fast loser. You've got to be on the cutting edge in order to be a part of the transformation. Which of those is right? What do you tell your clients? Listen, I think life is always in the middle and I think it is the strategy. At the same time, I would propose two by two approach where on the one hand to our previous conversation, you step back and ask yourself considering all these technologies out there. What is my mot and how defendable is my mode? And do I have an opportunity not just to protect myself, actually re invent myself, re invent my industry. And I think this is a very exciting place that I'm personally passionate about and work with a lot of clients on. And then number two is really going back to productivity use cases where you could clearly create a lot of value in a very short term based on the levers that you have today, that frankly, you could involve your organization. Now, it allows you to educate yourself. It allows you to educate the organization. It allows you to focus more on the risk side as well as on the tech side. So I would kind of do two of them but would not ignore posing and stepping back to find the remote. Good advice. Thank you for that. And, and, and, and how do you uh uh how do you come down on the the jobs debate? Is this going to eliminate jobs or is it simply going to change jobs? Well, I think the devil is in the details. Now we discussed, that's what you're here for, to give the details. So the details here are the debate around the future workforce was around an automation. So this thing spit it out. Number one, number two, you really need to go ro by role and understand the implications and put it slightly in the broader context in a sense. Number one, it creates a lot of new content and new opportunities and as we know, the creativity demand is elastic, so it creates more opportunities. On the one hand, on the other hand, it also able to solve to our previous, to your previous question. Quite a few mon questions, aging questions and a few other things if you ask me, I don't think there is a debate how many human beings or jobs it's going to eliminate. I think it anything is going to augment what we have today and coming back to see all the questions that we asked every day is really what does it mean for the future of your workforce? What type of talent you would like to hire? But more importantly, how do you educate the rest of the organization? Because you and I grew at the space where if I'm not able to explain you something in a few minutes, I'm probably not competent enough. And the underlying assumption here that you and I speak the same language that might not necessarily be the case in A I. And therefore you need to teach the organization the common language to be able to understand A I and it's opportunity. Yeah. Last quick question. Uh For the last few decades, I think it's, most people would agree that technology has tended to exacerbate decision uh divisions in society between uh the, the skilled and the unskilled workers. Some think this may be different that this may empower the bottom half of the distribution, less skilled workers, give them more superpowers. What's your view? Well, I think we're at the beginning of age of creativity because creativity is based on inspiration, inspiration. And many times based on the research, if you use generative A bi you could fast forward the research and unlock potential pretty much of every individual while eliminating some of the routine parts of what we are doing today. So if anything, I think it could close the gap and allow each one of us to be creative of our own destiny. Fascinating. Well, we'll have you back in a few years and see how it's playing out Alexander. Thank you very much. Thank you very much.