Eric Kutcher, Senior Partner and North America Chair, McKinsey & Company
Zig Serafin, Chief Executive Officer, Qualtrics
In conversation with: Diane Brady, Fortune
Zig Serafin, Chief Executive Officer, Qualtrics
In conversation with: Diane Brady, Fortune
Category
🤖
TechTranscript
00:00Thank you very much everybody so we're going to be talking about infrastructure and so much more. Nice to see you both. By the way Zig is a former musician. So accordion and sax just to kind of level set.
00:14Here we are. Here we are. He's going to play right. So you know I want to start. Actually you get into rooms. We don't get to go talking to CEOs. So before we get into the deeper dives on what it takes to get to the next stage of the revolution.
00:30What are CEOs talking about. Where are we right now. And you know give us some sense of just where we are in this evolution. U.S. and globally. I'm going to start with you. 60 seconds or less.
00:43I think CEOs woke up 18 months ago. This is real. They woke up 18 months ago. This is real. The same time I think they're extraordinarily frustrated by the lack of return on investment they've gotten so far.
00:54But I think they're starting to see these autonomous agents as a mechanism to begin to rewire and rethink and reimagine their business. So the business model transformation as opposed to some of the productivity gains and such.
01:09I think it'll be both. I mean I think I think what you're going to see right now is the opportunities to completely reimagine the way customer service is done. Right. Or I mean we can go through examples but you only give me 60 seconds.
01:20So true. So. So I think they're starting to say wait a minute. There is entire bits of what I do today that I can do completely differently with a much better experience.
01:31OK. I'm going to pull on that thread. So you know I get the privilege to be able to work with CEOs CMO CEO CROs all around the world different industries many of which many of you here representing.
01:43And there's a new framework that's underway for getting out of the boxes that we've been living in. If you think about the app based boxes that we live in today we're the ones that are providing the input.
01:54We're the ones that are actually the producers of the output that takes place. You know and so they've been neatly built as CRM applications website applications in the future will be one where that entire framework goes away and there's a new framework where you will have an agent based world where the speed at which you
02:12actually end up leveraging an autonomous engine to get something completed will make things very different. And so this isn't just about adding Windows dressing to the to the to the existing applications that we have today.
02:25Right. It's not about adding an AI feature to the world that we live in today. It's really stepping back and saying over the next two three four years how will someone engage with my airline.
02:35How do you understand that customer better. How do you actually create a faster way of enabling them in a more personal way as a result of that. That's a huge theme. Now along with that is responsible. AI is a major theme.
02:49Just like we have measures in most companies and organizations today where you have security controls and a policy around how you run security you will do the same thing for how AI uses the data.
03:01How you actually create trust. How do you actually engage with people in an honest way. Get permission to be able to create increasingly closer personal connections at scale across the way that your business shows up working with the customers and employees that ultimately end up enabling your business.
03:16I like can I pick up. Of course you can. You you talked about kind of this. I said reimagine. You know you talked a bit about the way people are going to think through customer experience.
03:24You made a huge announcement earlier in the year that I happened to get to see a little bit which if you think about introducing a new product promoting. I'm not promoting. I just saw it. It really just really threw it really. It was really quite really quite remarkable because today if you're going to you would go do a whole bunch of customer research right unless you're you know unless you're Steve Jobs you would go do a bunch of customer research you'd figure out what now you literally can type into your AI engine and because you have this incredible history of every interaction point you know what
03:54individual customers at very micro segments and so that information was never existing before. Yeah yeah. So that that capability essentially is called edge and it actually helps companies to be able to tap into knowledge about what's happening in their industry across industry and get instant answers to some of the biggest questions that might be on people's minds but also be able to get instant answers to the micro questions and decisions have to be made at a store level.
04:22Let's say that you're KFC brands for example and you're operating in a far region here. We have KFC here in the audience or if you're in the restaurant industry. What is it that your store manager needs to know about that took place yesterday across your competition and what customers care about in order to be able to inform the way that you end up showing up even the next day and of course at a bigger picture level is what trends do I need to pay attention to.
04:46And so being able to synthesize that type of data prior to the emergence of where AI is taking things would have been very expensive and very time consumptive right. Now you're moving into a world where the speed of getting that type of decision making improves and accelerates.
05:01So I think what we want to achieve in this panel and I'm glad we're addressing the audience because we will take a question or two is is both the potential of AI and the limitations. And of course one of the potential limitations is that infrastructure problem or issue. So let's let's let's just sort of do a bit of a lightning round here because I'd like to know how you'd prioritize addressing these challenges. Let's talk about the powering part the infrastructure data centers. How big an issue is that huge issue.
05:31Right. If you actually look at the amount of you heard the number from us before you look at all the growth in the data centers around AI. You don't actually have two issues. One is you got to build the data centers or actually don't have enough people to build them. Second you got to power the data centers. We don't have enough power. I actually think the power will come from frankly the big hyperscalers and others making massive investments.
05:52But that still takes time. It'll it'll be surmountable challenge. How about you Zeke. How do you think about it.
05:57I think about it in terms of two two two major lenses. Number one is you have the hyperscaler AGI people that are building these massive models that will require an immense amount of compute power. They have to be close to the energy sources. They've got to find new forms of sources of energy.
06:15And then there is task specific AI where you actually start to build fine tuned models. Ones that you know probably half of you in this audience will end up building and enriching and using a variety of techniques to be able to say it's purpose built specific to this particular task and that will be highly efficient.
06:34It'll be focused on a data set that is specific to what you do in your business and it would be cost effective as well. And so there's idea of actually building sustainable scalable AI is really a combination of those two major lines and we have the possibility of a trade war. I know we just had the China panel and obviously that's going to impact what happens with AI. Let's talk about the regulatory front since we're at a global event.
07:00Obviously we've got what's going to happen on the U.S. front with regulation which we've got some indication of. And then globally how much is that going to determine what this next stage looks like.
07:15Look the I think every every geography is going to do it differently. I actually think the U.S. will be much more about a framework for which to operate on and I'll actually put more of the responsibility back on the companies themselves.
07:26And I actually think that is not a bad model. I think Europe runs the risk frankly of overregulating and if they overregulate will they stop. Will they stymie the innovation if you stymie innovation. What happens from the economic benefits.
07:39What do you anticipate in terms of policy or regulation. Two things. Number one is I think in the background we have to really realize that we are entering into a golden age of AI.
07:52OK. You know the nature of where AI is headed is probably one of the most amazing technology innovation moments on the history of humanity.
08:04And and if you overregulate it if you over policize it to begin with you will end up stymieing the innovation opportunity. So it's more important. Eric and I were talking about this earlier that you create frameworks guiding points.
08:19Right. But ones that are aimed towards catalyzing the innovation that can take place. Right. And the second thing that's interesting is in that same time frame I think around the world we're seeing a time where governments are looking to get closer to understanding what's truly on the minds of the citizens and the residents that actually are part of the communities that they live in.
08:41Right. And we're seeing this at the micro level in the United States. Most recently if you look at the county level statistics that came about in the most recent election. And so what we're finding is that experience is essentially the currency of trust.
08:56And if you want to build trust you have to get closer and closer to the people that you serve. This is happening in the private sector. Right. If you think about the companies or the brands that you love and the ones you get closest to the feel understood you feel like you trust them.
09:09They're consistent. The same is true for governments. And so AI actually creates an opportunity to use the technology to actually be able to at speed understand how to get much closer to the people that are looking to participate in different programs that are those programs showing up in the way that you'd want them to show up once you deploy the funding. Right.
09:27So I'm going to ask you if you have any questions just raise your hand. We've got Mike runners here and I'm not sure if we do have any questions but as we before I turn to that I want to ask you who the net beneficiaries are going to be. We've seen the big tech companies obviously do well. You are a tech player and a lot of people want are concerned about the startup ecosystem and how that innovation flows from here. What do you see.
09:51I think we will be I think individuals will be the great beneficiary. I just again one of the big opportunities is going to be around customer experience. And if you think about things that are miserable experiences today we can completely solve that going forward. I mentioned earlier this Salesforce announced that their big conference this you know how many we've all this age of work but we all go through the experience where you dial a customer service number and it goes through 18 different ways to make sure you never get an answer.
10:17And now what they have is a voice at the other end that's completely done through agents and it walks you through to get you know what you need in a matter of seconds. And I just think this is early days of this but you start to go through all these friction points that I don't think will be friction points 18 months from now two years from now.
10:35What do you think. You know a couple of things. One analogy that comes to mind is is the interaction we have today with maps you pull out your phone you pull up a map you want to go to a location. We're still responsible for the interpretation of that right. And make sure the inputs right in the near future you know the map will be there in the background because the car the device is actually taking over and it autonomously gets you to you know where you need to go from point A to point B.
11:02That is the magnitude of opportunity and change that's underway across every single industry that will take place. And so you reimagine the way that we've done things right as a result. Now that said we also don't see a world where there's going to be just a few behemoths that ultimately end up creating these systems.
11:20And in fact what we envision is an ecosystem where people will innovate they'll build fine tune models they'll build agents that are specific for a certain set of business tasks or outcomes. And that will create a whole new world of opportunity that will take place. You know much like what took place in the Industrial Revolution. We'll see a similar effect. I see a question from Bob Nardelli here if you could just go in the middle.
11:42Thanks Diane. This is a great panel. Let me ask you a question. Do you see any similarities between A.I. and electric vehicles because there was the same excitement and euphoria about EVs the mandate that came down. It was range anxiety and it ended up turning charging anxiety right. No infrastructure. And you mentioned the point we've got to build a lot of things that are going to be able to be used in the real world and we've got to build a lot of things that are going to be able to be used in the real world and we've got to build a lot of things that are going to be used in the real world and we've got to build a lot of things that are going to be used in the real world and we've got to build a lot of things that are going to be used in the real world and we've got to build a lot of things that are going to be used in the real world and we've got to build a lot of things that are going to be used in the real world and we've got to build a lot
12:12of data centers and then we got to be able to supply energy to that. Now in Atlanta you know we just built the last nuclear plant Vogel years over overdue and billions of dollars.
12:24So I didn't hear any of you mentioned about small mobile nuclear reactors where you actually control your own destiny at the data center. Yeah I know some of Microsoft and some of the other are still out looking for a place to plant a data center.
12:38How do you how do you generally think about is energy going to going to throttle the rate with which you introduce a former CEO of Chrysler I should mention by the way that knows what I'll respond real quickly and you can talk specifically to nuclear and give what you're what you're saying.
12:54I the old saying that says we often end up overestimating what happens in the short run and we end up significantly underestimating what happens in the long run I think is extremely true here. I think that the capacity
13:07and level of innovation that I will bring it will take a long time for what those new scenarios can be to start to come alive. And if you look back at the Industrial Revolution back to 1850s 1860s when you know the Bessemer steel process was invented and how long it took for people to realize that steel could now actually help you build skyscrapers took many many decades later.
13:29Right. It was initially used in railroads and I think we are in a very similar stage at this point where people will see you'll see a I improve substantially but the timelines will be significant compressed you'll go from five year inflection points to three year inflection points.
13:43OK. What do you think. So I agree with that. In some ways the Internet was it was a very real example. That's right. The first Internet boom you know happened but didn't really ever deliver the promise and then it took time until we got there.
13:54I think we're in that but I don't think it'll take as long just by virtue of what we can do today and the availability. This question of power is a very real issue. Like it's a very real issue. The question of nuclear. The reality is we can't build even the small nuclear reactors fast enough. Right. It's going to take. So I do think that is part of the solution but I don't think that's part of the solution in the next five to seven years. I think we've got to think about alternatives.
14:18And frankly I think that means we are going to end up with kind of more natural gas and more solar and more. You're going to have more of the available technologies that are out there today that we're going to have to use as we start building out these data centers you know in the near term. But it also presses the need for more efficient ways of using AI and how you end up innovating that. So building smaller models specific to specific tasks right. You don't need all that data. You can be a little more depending upon what you're trying to do. And that's already happening. I think it's starting to happen. I see we're at the two minute point. I want to give each of you a chance to say a few words about what you're trying to do and what you're going to be doing in the next five to seven years.
14:48I want to give you a chance to give some closing thoughts in terms of root us in the moment but also give us a sense of both what's exciting concern. I mean help us see around the corner a bit as to what we'll be talking about this time next year. I'm going to start with using. Sure. I think first off we're at a stage where companies probably is not one CEO I've met with over the last year who hasn't said I want to use AI to get closer to my customers and to be able to scale the rate at which we end up
15:18differentiating ourselves in being becoming more personal more connected more convenient in time. Right. And that's a major theme. That said half of you here in this audience are probably building some kind of an AI capability one way or the other. Half of you will be consumers and you'll be using it right with that. And in most recent research that we had is there's only about 20 to 25 percent of consumers that actually trust the companies that are building the AI to actually make the AI perform as it's
15:47actually being promised and or be able to trust with how the AI is being used to facilitate them. Right. So being able to approach the problem with building trust where experience is the new currency that actually ends up driving how people will engage with your company and your application is going to be core you know a core essence in this.
16:06And I think that is in the context of the fact that a lot of people are just wary of all the hype. A lot of there'll be a big hype cycle that will take place and that we're in the midst of it. Right. Lots of claims promises the timing for that what that horizon actually looks like. I want to give also a chance for you to give some. I think you're down to a haiku or a limerick at this point. We discussed this. I'm just not that talented. I don't I think the technology will not surprise us or disappoint us. I think we're going to see the technology continue. I think the thing that most organizations are going to struggle with is talent.
16:36I mean this is going to be the biggest change that talent has to begin to understand and think about how they do their jobs differently each and every day and to reskill generations of individuals. I think that's the challenge we're going to face. Please join me in thanking our panelists. I know we'll just thank you again.