This is the McKinsey Podcast, where we help you make sense out of our world's toughest business challenges.
Welcome to The Show.
I'm Lucia Riley and I'm Roberta Fissara.
Today, why people aren't adopting Gen A .I.
more quickly. Despite the buzz, coming up an episode from our technology podcast at the Edge.
It features guest host McKinsey Senior Partner Brian Gregg with three tech visionaries, one from Mayfield Fund, another from Big Bass in labs, and the former director of trust and operations at AirBnB.
Speaking of J &AI Lusha, it's definitely having a huge impact on how companies are thinking about talent.
If more tasks can be automated, then what's the right mix of skills and roles that companies need to compete?
People should check out our recently published article about strategic workforce planning to learn more.
And when it comes to strategy around clean power, companies are rethinking their approach, especially as data centers are expanding quickly.
To learn more, check out our article about clean power.
And now, let's hear more about the topic of the Millennium Gen AI and why adoption isn't happening faster.
We have three tech visionaries with us today.
and being chata, Caram Presade, and Naaba, the energy.
Thank you for being with us for our audience.
Why don't we get started with just a quick introduction on your quick experience?
30 seconds or less, Navin.
Get a started. Navin Chata, managing partner at me field, early stage fund, investing up and down the AI stack, what we are calling cognition as a service.
I'm Caram Presade, currently the CEO of a small startup called Big Business Labs.
But before that, I was at a product and engineering at Nextdoor and before that I used to run consumer ads and subscriptions at LinkedIn.
Hi I'm Naba Banerjee.
I am the first female engineer in my family and in my last role I was the Head of Crossed and Safety at Airbnb before that 13 years at Walmart various roles.
Excited to be here!
Super excited to have the three of you here.
We have the investor, the product manager, the trust and safety thought leader.
It's so great to be with you all today.
I gotta get us started with a question that's rumbling around all of Silicon Valley and beyond is a trillion dollars later.
We're into this AI revolution, we call it.
Is it real? Is it a hoax?
Let's start with you, Naba.
Like, what is there to show for this?
And is this a real thing or is this just another headline?
I'm with you, Brian.
Like, when I first heard about Chad's GPT, I was one of the first people to go in there, get the app, pay for it even, and then I used it like crazy, and then we were going to replace all our customer support agents with AI.
Thank you for just going to change the world and record time.
And here we are, it's not happened yet.
So I am a little frustrated.
I'm hopeful about the future but it's taking too long.
So AI, I think, has a bunch of different parts.
So look, when I was at LinkedIn and then at Nextdoor, we used AI to rank your feed.
So AI is already having impact.
Massive business impact.
The big new thing is Genn AI, which is what's happened recently.
And look, it's been a year and a half two years, like you've got to give it some time before it starts to get adoption.
people have to learn how to use it and you have to build apps on top of it.
It's like, if you think about the iPhone, when it first came out, 2007.
So it was almost six years, six to eight years before you really got to a point where you can say, oh, iPhone is great, now iPhone is giant.
It's a huge business, trillion dollar company where 60 to 70 % of the revenues are primarily from that one device.
So I feel like we're on that same path, it's inevitable.
Over the next eight years, AI and AI agents will be the future.
But the hype, though, like iPhone didn't have this much hype, the first version of iPhone and then when the final version that surprising delighted us was amazing, whereas with A .I., there was so much hype, and now it feels like it's taking just too long.
What do you think, I mean?
The first thing is, A .I.
has been around for 60 -70 years before any of us were born.
Maybe our parents weren't born, right?
I look at it as an evolution.
And any time there is a new technology, and especially the one we're all interested in is Jenny AI, it's only two years old from the launch of chat GPT.
And I would say AI -based applications, there was more hype because the technology 10 years back wasn't there to be able to do things that were advertised.
10 years later, a lot of the things that we're talked about are going to hit reality and my belief is in semiconductors?
That's what Silicon Valley was built on with Moore's law, the Garden of Moore, the founder of Intel, every two years processors paid double.
I think every two years, the impact of JNIA is going to be four X.
Is this happening in two years time?
Four year time, what's the investor mind and balance view say?
So I think it depends, Brian, on the use case, If you look at consumer's and prosumers where on the other end you don't have bureaucracy buying this stuff, it's going to happen very, very quickly.
So if you look at the adoption of CHAD GPT 100 million monthly active users, months, right?
You go before that Instagram two, three years, Facebook six years.
Every time it's half, if you're going and selling to a enterprise IT, and there is a human buyer, that's where the friction will be slow, and it'll also depend upon are you taking human jobs in the enterprise, or are you filling jobs that humans don't want to do, can't do or there's a shortage of talent.
So I'm very bullish and very fast adoption in consumers, which we have seen, consumers, developers, and micro businesses.
And enterprise, there's friction.
There's an IT buyer, there's a chief legal officer, you took my privacy data, training can't be done.
Then the biggest issue is they want to host the data in -house, we do all the worker companies, then you ask them, what are your top two problems?
I don't have a business case.
The second question is I don't have talent to implement, But we give us a different view on from the front line queue and maybe a knob of them too as from front line, people in the business you're starting to come up right now, and how fast does this happening?
First, start up like me, it is now.
Like if you are looking at all of the tools we're using, everything's AI.
I'm probably using AI like 300 times a day easily.
And that's not just for like the coding side of it.
Build our logo, build our website, build our marketing material, build our customer support site.
everything is all AI first. I still think it's gonna take a long time for adoption because I think like any new technology, it's just gonna take time.
Google, when it first came out, Google search engines, also they used to all these things.
I don't know if you remember.
There used to be companies that you could call and then you would ask them to do the search for you and then they would do the search on Google and then give you the answer.
It was because, yes, because they didn't know how to say and , and, or in site colon and all this stuff and then it took years before people learned how to use it.
So I don't think that the tech is far away but I think the user adoption to understand how to engage with an agent and use it effectively to accomplish things is something that's just gonna take time.
So I think the solutions will be there but the adoption rate will be potentially lower.
Now, but take us back though, 2020 2021, right?
When you were coming to Airbnb in the trust and safety role, how to change from there to when you ended that stream of your career.
How did AI influence what you did and how you did it?
I will actually give the moral of the story before I give this story which is I think the biggest mistake we make is thinking about AI for the sake of AI.
What will never go away is the thing that humans really do well which is articulate the problem really well.
So when I joined as the head of Trust & Safety at Airbnb.
It was a really difficult problem like the world had gone into lockdown bars and hotels had shut down and teenagers were throwing parties on Airbnb and I remember just sitting down and not having a clue as to how do you even start?
But the thing that saved me was I got a group of cross functional experts including getting advice from police chiefs or our comms partners, our designers, our developers.
We knew we needed an intelligence that can keep up with the trends in the world.
That's when we built the first AI model, then rolled it out in America and then whole world.
And three parties, incidents are less than 55 % compared to when we started.
So that's something that I think we should never forget, that it's never AI for the sake of AI, but to solve truly problems with specific problems. Well, and maybe we flash forward a little bit, and Naveen, I'll start with you on this is, if you were to flash forward two, three years, let's say the adoption curve is what you just said it is.
What do institutions of 2028 look like?
Is it half machines, half people?
What's your view of that?
Yeah, so we believe every human is going to have a digital companion and we call it AIT Mayts.
And our strong belief is these AIT meds and humans will work together.
So that humans really work at their exponential potential, what I call it, human squared, what it means is, AI will have to do more than automating tasks.
an accelerating productivity.
So essentially we have to start thinking about how does AI augment human capabilities how does AI help me amplify my creativity?
And it's really a teammate.
It's not assisted intelligence with a co -pilot or it's not like go do this task.
How do we get better together?
So future organization will be digital workers along with human workers, the organization of the future will be a hybrid organization.
And the CEOs and executives who are in Dorset will get on the other side and people who don't will end up becoming dinosaurs.
This is the same that happened with the internet, with e -business, you don't have a mobile app.
You know where you go, right?
So that's what is gonna happen.
This is like a critical thing.
The agent or the team may sort of approach versus the supported approach And I'm a pretty strong believer in the agent approach. So like the best way to think about this, I use this small metaphor.
If you were going to write a book today, most people would open up Google Docs, or open up word they would start typing their, their book.
And then maybe you're getting like, spell check and grammar check to kind of help you.
So like, you know, there's a little AI that's kind of helping you.
And then you'd send it off to an editor and then the editor would review it.
They'd give you sort of feedback.
You'd update your document and then you'd ship your book.
I think you should think about the agentic world like you're going to have a ghost writer.
Okay. So if you're going to tell your story you go to the ghost writer, the ghost writer writes the book.
You then provide editorial feedback on whether or not the book is good or not, and which parts to fix.
And you can work with the editor to make it more creative.
Oh, this part doesn't sound as interesting.
And so I think this idea that you're really going to have a teammate, somebody who's going to do the work.
And you're going to guide it and poke at it and, you know, give it direction a little bit here and there.
That's the future. Yeah, it's collaboration, right?
Like that's what we do all do.
I love that both navine and Kyren, the only push that I have is as an operator, every year we would go to our CFO saying, I don't have enough money.
I want more resources, I want more engineers to do the work.
And you're like, but now, but now you have AI assistants or you have AI copilot's like, why do you need people anymore?
But if we go back and say actually now AI will help our engineers be even more productive the record. So that means I have to keep all of the engineers and I have to pay for AI so you want to make me to double the spend and how much time will it take for you to be double productive become back to that question.
I imagine I want to imagine this beautiful world where.
Yes. I don't be amazed.
We have to start from the fringes, right?
Go after jobs, then just count higher talent.
So let's look at DevOps engineers.
Like you look at IT ops engineers.
You look at security engineers.
You look at chip engineers.
Same thing which happened with IT outsourcing or manufacturing outsourcing.
Don't take the high end of the knowledge work and replace those.
Go to the fringes. Second, go after things humans are not good at.
It's sifting through a case law.
Preparing for litigation.
Can you boil it down to 90 % knows irrelevant?
10 % is what you give to them.
I think the other thing is if the CFO is not using AI, they can't understand how to, what it makes, what it means.
I create AI agents with different personalities and I have conversations with them to then say, that is my teammates that I'm going to interact with.
I literally did this.
We could see. Yeah, we recently tried to raise funding.
and I needed legal advice, I was like, okay, I'm gonna set up five different lawyers and I'm gonna say they're all going to talk to me.
I uploaded the contract and we diagnosed the contract and looked at what the pros and cons were.
I had them argue with each other.
Amazing. And so things that, I think people, this is what I mean, I think we're in those early Google days where people were like, you can do that.
That doesn't even sound real, is that a thing?
And I think your CFO just doesn't know yet.
It's a one way to solve the budget problem is my advice to every teammate, or people who are on agent tech architectures is, don't charge for the number of hours.
Don't charge per seat.
Charge for the work you do and the outcome you create.
Right. Please change it.
This is a model. Completely changing the model.
And I think it's the same which happened with perpetual license, pay me up from for five years, no. The next company comes pay me monthly.
Yeah. Right. and the cloud compute happens pay me as you use like you do for electricity.
So get these digital workers, they're off most of the time.
The answer calls, they reconcile AR, they file tax returns, then pay them.
So I think once the tech is getting there, but the workflow isn't there because they need enough practice, sure to get better.
And then in an enterprise setting, there's one more thing I also don't like is it's on close data.
So the amount of training that is required sometimes on open data on the internet is much easier to create a scalable service.
But SAP unused, I have custom data.
So it's complicated.
That's why you have to go on the fringes where, but I think it requires business model innovation.
Make it so easy. Yeah, that's pretty easy.
Understood. The mapping shift I think is similar to Uber.
I think if you originally wanted to have a driver, you had to make enough money to have a driver and pay them full time.
on a regular cadence once a year, once a month, whatever.
Then we're made it so you can easily have access.
But it doesn't mean everybody got rid of their car.
In a world where you have half machines, half humans, what is the leadership team of tomorrow look like?
How does the CEO? How does her, his team operate in that hybrid world?
I think, it will actually take of you a lot of fear associated with leadership.
I think people who want to start their own companies are people who want to lead companies Or be a senior leader.
They think they have to be this person of exceptional talent who has to have this very creative vision, has to have all the best decisions all the time.
They will be able to use AI to say, simulate these five scenarios for me and give me all of my risk versus benefit numbers.
Help me understand if I'm going to get sued or not by exactly what you are doing here and come up with all of these creative ideas and challenge each other.
I love that idea that suddenly all the things that because everyone can't be exceptional at everything, that everyone is exceptional at at least one thing.
But those other areas of your personality that maybe kept you behind, now you will push forward, so we will probably see a lot more leaders emerge, but on the flip side, then it'll get harder to distinguish yourself, because suddenly, it's an equalizer.
Everyone has the same resources available, so that's the conundrum that I'm not a fortune teller, I'm very excited to see.
Karen, let's put it on the table.
A lot of today's CEOs follow a certain track, whether being an MBA, a graduate degree, got a job usually in a commercial function marking sales sometimes product, and worked their way up.
What's this CEO of the future look like?
Is it the same pathway with a few tweaks?
And you're playing the role right now?
I think it's the same pathway, but in more than a few tweaks.
So I think, like, look, the part of what you do is you get into larger and larger leadership roles.
Is you get really effective at understanding strategically where you want to go and then delegating tasks.
And in an analytic world, you will be able to choose which tasks delegate to an employee versus delegating to an agent.
But you still need somebody who's setting the strategy.
I think what will continue to be an even more important skill is communication.
How effectively and concisely can you convey what you're trying to accomplish to a person versus an agent.
And what happens is it kind of permeates through an organization is that it typically like dissolves.
And in an agentic world, you're gonna be able to maintain fidelity, going from agent to agent to agent.
It's trying to accomplish things.
So I think the more effective you are, the game of telephone will be more precise.
You have to get really, really clear on being able to predict where the future is going and guide strategy more effectively.
So the whole like, I'm going to just A B, test it.
Stallone is going to go lower.
And if you can you see this across a whole portfolio of companies, what would you agree with us?
Is there a different?
Yeah, version of the CEO. I think the CEO, I look at it as right.
They always have to be raising money because without money, you can't do anything.
And secondly, they're in the business of mobilizing resources.
This time it won't just be human talent, It will also be AI teammates.
And then you have to make decision, but the smart CEOs like athletes surround themselves with coaches and this time around, I'm gonna have a lot of digital coaches who can improve my serve, CEOs have a tough time giving feedback, I'll have a candor coach. They might be afraid of speaking the best ones, demonstrate vulnerability, others don't.
They have to maintain a persona, but with a digital team mate.
It's all confidential.
Only you and the team mate sees it.
So I think they need to be, I would say, an AI native CEOs.
And my only input to them is, get somebody with a fresh mind as they are chief of staff.
Now the question is, is it a digital team man or a human team mate?
Maybe it's a combination, first principles thinking to tell them how to operate this puzzle.
my view. That's a digital team mate.
That's good. I don't know.
It should be. I should be.
I mean, it's if you just, it's too full of stuff.
If you, if you look right now at like, what is the biggest adoption for the AI stuff beyond like, touch, P Airport, the other ones, character AI, replica are all of these things that are effectively like, it's like high trials.
They effectively are people who are like, Oh, I want a safe.
Environment for me to communicate with, I think a lot People trust weirdly enough people keep talking about, like, I don't know if I trust AI, but the number one use case that seems to be working is the one where they have to trust the AI, just like insane.
So, you brought it up and you know I'll start with you and then chat enough to chat.
If you can go to, start with you and then not wanna go to you.
If we're talking about 2028 where half the jobs are done by these teammates, these digital agencies, ai agents and chiefs of staff, what is the downside effect on humanity, on in the employee base on society?" Yeah.
So I think humans are smart.
I look at AI is yet another horse.
It's yet another tool.
Humans will figure out how to ride it, the way that it PC's mobile.
We'll just get better.
And when this productivity is more revenues, more profitability, more jobs get created.
So essentially when GDP growth happens, right?
Like, keep up so some of it will be.
AIT made some sound like very bullish everywhere.
It was very positive.
Yeah, every time Attack Waver happens, right?
Like, humans win you know what?
If that's not going to happen, they'll never pay a bill for this right?
Like, so my thing is short -term pain, long -run.
It'll be all natural impact for him.
That is the same way, the great equalizer offshoring happened people taught india will take away all the jobs, the U .S. God richer and richer with globalization.
I feel like I have to balance that view with What's happening with these governments?
I am an optimist. I am too.
I am too. But you are so much of an optimist that I almost have to go on the other side.
Otherwise, I won't be in this business.
You know, for 13 years, I was at e -commerce.
Like I worked on growth and helping people buy.
And then the trust in safety world exposed me to another part of humanity that at times, I wish I hadn't seen.
And I know different marketplaces, dating sites, are trying to create an environment where humans can meet each other.
Now after COVID so many people are almost meeting with a first -time digitally.
And if it's always scary for humans, like stranger danger is still considered to be one of the top fears that prospective hosts and Airbnb have. About 60 % of prospective good guests say that they are scared of being scammed.
It's not true. Like, very few incidents actually happen, but this is a fear that humans fundamentally have. and with synthetically generated humans through AI.
It's so easy to recreate voices of people, digital twins, fake IDs.
We are seeing the way we have typically kept communities safe, trust and safety and risk teams what they have done.
All those defenses are failing.
We are not ready for the world that is coming.
There's also a lot of bias in the data that has been used to train these humans, and so, yes, it feels like red tape in the privacy team and the anti -discrimination teams come and say, you cannot launch this model.
We have to watch the data.
And I used to push against these teams, but I realized that it is happening.
We should be solving for this, but we need to go in with eyes wide open.
That the same AI is in the hands of good actors and bad actors.
And so we have to constantly think about the two sides of the coin.
Well, with that, I wanna thank you all, Nevin.
Now, Lucilleryn, thank you for being with us.
Thank you very much. Thank you.
Thanks so much for listening to the Mackenzie podcast. I'm Lucia Radioli and I'm Roberta Facaro.
Find us on Mackenzie .com.
We'll have a transcript of this episode up shortly and download the Mackenzie Insight app where you can find this podcast in other helpful content, updated daily.
If you enjoyed the show, we'd love for you to leave a rating and a review.
We'll see you in two weeks.
