The Shift Code
PMI CEO Pierre Le Manh takes listeners inside real stories of organizational transformation from sectors and regions across the globe. Join us for candid conversations with top leaders in transformation as they give a behind-the-scenes look into the strategic and digital innovation driving their journeys, their lessons learned and the professional skills needed for success.
The Shift Code
450 to 5: How WPP Rewired for the Future of Creative Intelligence
In this episode of The Shift Code Podcast, host Pierre Le Manh is joined by Stephan Pretorius, Chief Technology Officer at WPP, to share an unfiltered look at how one of the world's largest marketing groups is using AI to reshape strategy, creative work, and even its business model.
What You’ll Learn:
- How to drive enterprise-wide transformation through technology platforms while maintaining psychological safety and entrepreneurial spirit
- The three-phase approach to transforming highly federated organizations
- Why successful AI implementation requires multidisciplinary teams and M-shaped professionals who combine broad expertise with deep specializations
- How to evolve from time-based billing to outcome-based commercial models in professional services
- Why humanities education and critical thinking are crucial competitive advantages in the AI era
Stephan Pretorius was appointed as WPP’s first Chief Technology Officer in 2018. He leads the company’s AI and product strategy and currently serves as interim Chair of AKQA. Previously, he was UK Group CEO and Global CTO of Wunderman, joining the agency in 2016. With a unique background combining law and technology, Stephan has been instrumental in transforming WPP from a federation of 450 agencies into five integrated global organizations through technological innovation. As the architect of WPP Open, the company's proprietary AI-powered platform, he has led WPP’s $400 million annual investment in internal product development, revolutionizing how the organization delivers creative and marketing services.
Episode Resources:
STEPHAN PRETORIUS
You need smaller teams of more expert people, but what you also need is you need more multidisciplinary teams. A great strategist, a great creative, a great CRM person, a great product design person, all working together using the same tools, can move mountains.
MUSICAL TRANSITION
PIERRE LE MANH
Hi, Stephan.
STEPHAN PRETORIUS
Good morning, Pierre. Yes, lovely to be here.
PIERRE LE MANH
Great to have you.
STEPHAN PRETORIUS
Yes, lovely to be here.
PIERRE LE MANH
So, how is Cannes for you?
STEPHAN PRETORIUS
It's excellent. I mean, it's always really hard work. You plan it months in advance, you stack your meetings, you get triple-booked for every slot, and then you just work through it as quickly as you can, as long as you can.
PIERRE LE MANH
So everybody's talking about AI here. I thought this year would be already gone, but it's not a fan, right?
STEPHAN PRETORIUS
Yeah, I'm not sure where it would go. Look, AI has been a theme at Cannes for a couple of years. Two years ago, we had Jensen Huang on stage here with us at Cannes. At that stage, no one knew who NVIDIA was at Cannes. You wouldn't believe it, right? It was pre their kind of massive market cap rise. And at that time, Jensen said to our CEO, Mark Read, on stage, he said, WPP will be an AI company. And he wasn't just saying this because he was provoking. He was saying this because of all the things we're doing with him and all the innovation we're doing in this space. So I think it was prescient. And two years on, I think the whole industry is beginning to realize the full impact of AI, not only on marketing and advertising as a category, but on all knowledge work, all industries where managing, processing, information, knowledge, making decisions is important.
PIERRE LE MANH
Yeah, this is very important for project professionals, as you can imagine. But let's talk a bit about your business at WPP. So how does AI affect WPP?
STEPHAN PRETORIUS
So WPP's business essentially is a professional service category that exists to connect brands to consumers. And I know that sounds super simplistic.
PIERRE LE MANH
That makes a lot of sense.
STEPHAN PRETORIUS
Yeah. And in between, there's a lot of complexity.
PIERRE LE MANH
Does it also connect consumers together, brands together.
STEPHAN PRETORIUS
Yes, of course. It's the entire ecosystem. But in the simplest terms, it's to connect brands to consumers in order to help our customers to grow most effectively. And more functionally, our job is to help our customers to invest their marketing investment in the most effective way in order to achieve that growth. Now, in that most effectively are a lot of choices. You can invest in product. You can invest in customer experience. You can invest in owned channels. You can invest in marketplaces, social, influencer, paid media, CRM, all kinds of things, right? And all these decisions are ultimately mutually competitive. And when you have an ecosystem optimization problem like that, as you can imagine, it lends itself really well to algorithms and AI. And when you add to that the fact that what we put through the pipes, what we let flow through those algorithms is essentially content. With Giant of AI, you now have two layers of operation. You have the investment optimization decision, but you also have the impact on the content generation and how human creative people work with AI in order to make that content.
PIERRE LE MANH
So basically what you're saying is that a lot of creative work is going to be or is already automated by AI.
STEPHAN PRETORIUS
Let me qualify that. So I think when people hear questions like that, the first thing that jumps into their mind is someone is just going to sit there and say, make me an ad for KitKat and the ad will be produced and sent out on the internet and sales will grow and everything will be fine, right? That's not exactly how it's working. How it's working is that we are using AI at every step of the marketing and creative supply chain in order to make great work that works for clients. Now, what I mean by that is that the first step is to define the problem. The first step is to see, you know, where does the brand sit in its ecosystem against competitors? How do consumers perceive the brand? You know, how do you model those audiences into AI agents so you can assess work? How do you come up with better strategies? Then the second step is to say, well, how do we develop communication ideas? How do we test those ideas? How do we then convert them into initial concepts before we take it to production? And I'd say what really happened in the last 12 months is that we've gone further down that chain, you know, that supply chain than I thought we would a year ago. Yeah, with AI.
PIERRE LE MANH
So how do you practically use it? Does AI generate tons of ideas or big ideas that then you can test automatically? How does it work?
STEPHAN PRETORIUS
Yeah, exactly. So we've built a platform called WPP Open, and it's our own proprietary secure environment where all our workflows, our domain knowledge, our data about brands, audiences, etc., is encoded into the system. And we have a whole bunch of tools for strategy, for creative ideation, etc. And so up to last year, I would have said we can use AI to come up with really great ideas. So fully formed strategies and ideation concepts, visual concepts. But then you would still take those and then you would go and produce them in traditional ways. In the last few months, we have started producing full broadcast quality, HDTV quality video assets that are being flighted using only AI.
PIERRE LE MANH
That's incredible.
STEPHAN PRETORIUS
Yeah. So we produced a campaign for an AI data center company called Weave in the US recently. They had a very short space of time to redesign their brand, create new brand assets, and then also make a TV ad that was flighted on CNBC and Bloomberg before their IPO. And then there was no time to actually shoot something. So we made the entire thing with AI.
PIERRE LE MANH
Does it show? Like, can you see it?
STEPHAN PRETORIUS
No.
PIERRE LE MANH
And there are characters like human beings in the ad?
STEPHAN PRETORIUS
Lots of humans doing real things and working in fields, working in data centers, working in factories, working in offices. It's indistinguishable now. We've remade Super Bowl TV ads that were shot over many months for some of our clients as a sort of illustration. We created it with models like VO2 and VO3 from Google. And most people get the distinction wrong. Most people think that the original is AI, not the other way around.
PIERRE LE MANH
That's very interesting. So it compresses timelines. Does it compress casts, honestly?
STEPHAN PRETORIUS
Absolutely. It really does. I mean, the current state of the art is that you need to do lots of, particularly for video, because video is a very complex modality. You need to do many, many prompts in order to get, and complex prompts in order to get the shots and the, you know, the kind of the clips right. So, you know, for that Super Bowl ad that we remade, it was two producers working over a weekend, but they made 3,000 video clips that they then selected the best ones from session together and then edited the final product, right? Still, it was two people over a weekend.
PIERRE LE MANH
That's unbelievable. And then the multiple formats, that's now you need 30 seconds or 15 seconds and eight seconds, formats for specific platforms. You can do all of that very quickly, right? The scaling of that is underlievable.
STEPHAN PRETORIUS
So it definitely does compress timelines. I mean, I think what people need to understand is that it doesn't take away the need for skill, craft, storytelling.
PIERRE LE MANH
Sure. Strategy, understanding consumers. But how many people is that compared to what it was before?
STEPHAN PRETORIUS
I think ultimately, and this is sort of a general realization that I'm picking up in conversations here at Cannes this week as well, is that I think most people who are close to this are beginning to realize that the shape of the workforce is going to change. And in two ways. The one is you need smaller teams of more expert people, but what you also need is you need more multidisciplinary teams. So a great strategist, a great creative, a great, you know, CRA person, a great product design person, all working together in similar interface, you know, using the same tools can move mountains. So multidisciplinary is really important. And I think even within individuals, we talk a lot about M-shaped people now. So, you know, moving from T to M. Maybe you're going to elaborate. So the concept of, you know, think about kind of skill sets. You know, some people are generalists, so horizontal, the T bar at the top. Some people are specialists and they go very deep on one topic. And for a long time, people have started talking about the need for T-shaped talent.
PIERRE LE MANH
A few experts or some experts and some generalists.
STEPHAN PRETORIUS
Or actually people who are both things, you know, they understand the broad field and they can contextualize what they do, but they're expert in something. And what we're now beginning to talk about is the need for M-shaped people, which means generalists that have multiple specializations, you know, so people who are generalists about the entire category of marketing, but can go deep on media or PR or ad tech or AI or, you know, creative.
PIERRE LE MANH
And that's because they become so productive that they can afford being specialists of many fields and they integrate them.
STEPHAN PRETORIUS
Exactly. And I think it's because fundamentally what AI is doing to our processes, it's not just speeding up or in some cases automating individual parts of our workflow. What it's also doing is it's forcing us to conflate our workflow into a single process. So previously, and you've experienced this, I'm sure as a business leader, if you want to get something done in terms of marketing, you often had to brief multiple people in series and then you brief Kantar for the consumer research. You get the report back. Then you brief your creative agency to come up with a concept to address, you know, that research. You get the concept back. Then you brief a production company to go and make content to fly it across multiple channels. You get the back. Then you brief a digital product agency to go and build a website or an app that takes that content, puts it in experiences and then runs it. Then you need to brief your CRM team or in-house or otherwise to then run email marketing and everything else against that. So that was the fragmented, siloed, historical pipeline. What you can do with AI platforms now is that you can pose your business problem, give it all the knowledge that you have about your consumers, your audiences. You can encode those into AI agents. You can come up with ideas. You can come up with sample images. You can approve those. You can test them against the audiences. You can adapt those for multiple channels. You can then even push them into the channels and start getting the feedback loops for the response. You can do all of that in one workflow. That's ver
PIERRE LE MANH
interesting. So let's talk about WPP a little bit. WPP used to be basically a financial model, right? You would buy companies and have a very accretive system where the multiples, the valuation of WPP was higher than the valuation of the companies you were acquiring.
STEPHAN PRETORIUS
Exactly.
PIERRE LE MANH
And that's how Martin Sorrell built that group so successfully.
STEPHAN PRETORIUS
Genius financial model.
PIERRE LE MANH
So in this world, what does it mean for WPP? How do you feel WPP should be organized going forward?
STEPHAN PRETORIUS
It's transformational for us on every level, right? It's transformational for us in the sense that we can't be effective. We can't leverage our scale firstly if we don't have one platform and one technology stack to work on. So what I've been doing for the last seven years is to build that platform, WPP Open. The first iteration of it was really kind of to help us organize our different sort of service offerings into one kind of organized workflow and standardized way of working. But now, of course, with AI, it's given us the ability to really accelerate deploying AI across the business. So the need for common technology is the first thing. The second thing is that it's changing our commercial models. So it is fundamentally changing how we charge clients and what we charge them for. So we are moving from almost exclusively T&M, FT kind of models to more output and outcome based models. And in some cases, entirely outcome based. So these sort of mixes of saying to clients, well, it doesn't matter how much time we spend on it, but we're going to charge you for what we produce and ultimately how well that works. So that's
PIERRE LE MANH
very interesting because the history of advertising or the advertising industry is interesting from that standpoint, right? It used to be paying a margin on media buy.
STEPHAN PRETORIUS
Yeah.
PIERRE LE MANH
And then FT Plus+, cost plus. And so now you're feeling that with this system that you have, you can move away from that and go back to like consultants do, right? Charge on a certain outcome and then increase your margins.
STEPHAN PRETORIUS
Exactly. And there's sort of two things at play here. The one is that AI is helping us to kind of to produce a lot better work more faster and therefore effectively kind of create more margin opportunities. But the other thing that's happening is that when you have only one lever to pull, if you're only making the ads, you can't really control the outcomes. If you're only making the adaptations of the ads, you can't really impact the outcomes either. If you are simply buying the media, you also can't really control the full thing. And so the conflation of the service categories that AI is forcing is also allowing us to have conversations with clients to say, we'll do the whole thing for you end to end. We'll be your end to end partner that does strategy, creative, production, media, CRM, all of this for you. And because we are controlling the full decision set, like the full problem set, we're able to optimize that and be responsible for the outcomes.
PIERRE LE MANH
So our audience is made of many professionals who manage transformations. What you're describing here is moving from a very fragmented style organization to a very integrated organization.
STEPHAN PRETORIUS
Yeah.
PIERRE LE MANH
So how many companies do you have at WPP today? How many agencies? How many brands?
STEPHAN PRETORIUS
Seven years ago, we had 450 individual agency brands with their own P&Ls, their own CEOs, their own IT systems, their own ways of working.
PIERRE LE MANH
And that's not even accounting for the local countries, right? The local markets. You multiply 450 by whatever number of markets.
STEPHAN PRETORIUS
Exactly. Today, we have five.
PIERRE LE MANH
Wow. Okay.
STEPHAN PRETORIUS
We have simplified the business down to effectively five large global organizations. We have WPP media, our single media buying and planning business. We have Hogarth, which is our production platform. It's nearly 7,000 people that does production in many, many different offices and studios around the world. And then we have three large integrated creative and digital agencies. So Ogilvy, focusing more on advertising and brand building. VML, focused more on brand experience and customer experience. And AKQA, focusing more on innovation and future strategy. And so that's the portfolio. And it's been a gargantuan job on many levels. Emotional, people, leadership, transformation.
PIERRE LE MANH
So let's talk about that. How do you even start that?
STEPHAN PRETORIUS
It starts with leadership and a vision. And, you know, I think about transformation, and this is very much my own personal sort of, you know, way of understanding it and what I've learned through the process. But I think about transformation as being a bit like warfare, you know. And you have to be very good at propaganda, but you also have to be very good at trench warfare, right? So you need to say very simple, very clear things, very consistently, all the time, and repeatedly, many times a week. Literally never get tired of it. Just like say the same thing over and over and over many, many times.
PIERRE LE MANH
That's hard to do, right? It's very hard to do. I do this all the time, and I feel I'm boring.
STEPHAN PRETORIUS
You get very bored of yourself, but you have to do that. And then the second thing is that you can refresh it. And like once a year, you can refresh, you know, you can evolve the message, but, you know, you've got to be consistent. And then the second thing is you have to be good at the blocking and tackling of everyday change management. Get the early adopters on board. Get the coalition of the willing sort of to come with you. Make it infectious. Make it successful. And I think the most important thing is you have to lead through example. You've got to perform the change, if you know what I mean, right? So, I mean, in the AI space, it's no good me just talking about theory and talking about the, you know, the sort of the vision. I have to, and I do, I pull out my laptop. I start doing demos, and I show people how it works, and I show them what they can do with it. And hopefully I inspire them, and I inspire clients, and the thing gets momentum, right? And I've been very lucky because my partner CEO, Mark Read, is a, you know, at heart, a technologist. He loves technology. He's given us an enormous amount of scope to push hard and invest hard in this area. I mean, we're investing $400 million a year in internal product development.
PIERRE LE MANH
Yeah, out of what is your revenue right now?
STEPHAN PRETORIUS
Yeah, so it's about 3% of net sales. It's whether you think the percentage is right or not, it's a big budget, right? I mean, this is a very well-funded startup. Think about it- So, you know, we've had the resources and the space and the platform to invest in this heavily. And Mark himself, he does this without fail. He pulls out his laptop and he starts demoing our platform to clients. And in breakfast meetings.
PIERRE LE MANH
And he used to run some of the agency.
STEPHAN PRETORIUS
Exactly. Starting with visible hands-on leadership and basically the whole notion of show me, don't tell me is really powerful. I think then the other thing is you need to give people hope. It's a very underrated, for me, element of leadership is giving people hope. People say, oh, it's too soft. You know, hope is not a strategy. I mean, actually hope is a very important emotional and personal need, right? People need to feel hopeful about the future of their company, their careers, etc. And so the framing of how you communicate the change so that people feel hopeful is very important.
PIERRE LE MANH
So in the case of WPP, for instance, I'm assuming that with all these brands, founders of companies who had sold their business to WPP and so a lot of identity, emotional connection to their local brands and agencies, how did you give them hope? And not exactly the opposite. The sensation that everything is going away and it's now the big machine taking over.
STEPHAN PRETORIUS
You make the best ones the future leaders. I mean, WPP today is effectively led by founder, entrepreneur, CEOs, almost across the board. Almost all of us that are running the large companies.
PIERRE LE MANH
You sold your company at one point.
STEPHAN PRETORIUS
Exactly. I sold my business to WPP 12 years ago. Jon Cook sold his company. He runs VML to WPP. Brian Lesser runs Real Media. He's an entrepreneur who sold his business to us. I mean, so it's a model that can work and can be harnessed. I think the sort of fiction that people sell companies and then leave and do something else can do that if you're selling a product company or if you're buying a product company.
PIERRE LE MANH
But it's a people company.
STEPHAN PRETORIUS
You need the good people to serve. The hope is really important. I think practically, you just need to invest heavily in training. You have to invest so heavily in adoption and training. We started really early. We started in 2019 with a program called Our Future Ready Academies. And we started training people on demystifying AI. What is AI? What is generative AI? What are, at that time, what were GANs? What are, you know, all these kind of things. So we started super early. And of course, that allowed us to accelerate when all the generative AI stuff kicked off in 22 and 23. And then I think the most important thing ultimately is that you need to actually then show the results. You have to link all of this back to, okay, this is happening. This is what we're investing. And this is what it's doing to the outcomes. We are winning more. We are able to stay more relevant. We are able to price differently. You have to bring it all back to kind of concrete business results ultimately. I mean, otherwise, it's politics.
PIERRE LE MANH
Of course. So how do you manage the governance of this transformation? Do you have a PMO office? Do you have a TMO office? How do you do that?
STEPHAN PRETORIUS
In our case, what we did seven years ago is that we set up a central office of the CTO in WPP. And we didn't call it a PMO office or TMO office, but effectively, that's what it was.
PIERRE LE MANH
But a concept deal.
STEPHAN PRETORIUS
Exactly. And we sort of created a couple of work streams around partnerships, around, you know, around data, around product development. And then effectively started working with all the agencies. And as the business was simplifying, obviously, it became a lot easier to drive the change. I think the phases, and this might not be entirely generically applicable. I mean, this is maybe quite specific to our experience. But, you know, in such a hyper-federated, hyper-fragmented ecosystem, we had to go through a few phases of transformation because if we went too hard, too fast, it just wouldn't have sucked. We would have broken things or we would have broken trust. So the phases were roughly, phase one was build community and trust. So we said, all the developers, all the technologists across the agencies, oh, fantastic, you exist. I'm a friend. I look like you. I speak like you.
PIERRE LE MANH
So it's interesting. You started with stakeholder management, basically.
STEPHAN PRETORIUS
Exactly. We started with building a tribe. And we call it the WPP creative Technology Council. And we brought all these people together. We brought them physically together. We introduced them to each other. You know, we started, you know, like doing things together. That was phase one. Phase two was, I said to, once I understood everything that people were working on and all the projects that were building and tools that were building, I gave them a gift. The gift was, I will accelerate your project with central funds if you do this in a certain way.
PIERRE LE MANH
And so what was the way?
STEPHAN PRETORIUS
Well, the way was a business plan, a roadmap, KPIs for the product.
PIERRE LE MANH
So proper product management.
STEPHAN PRETORIUS
Proper product management, right? So we professionalized product management by funding acceleration. And that worked really well. People, there were a few kind of like red accelerators. People bought fantastic products, started monetizing them. The second year we said, you can only get this funding again if you now not only build the product in a professional way, but you also build it on a common platform stack. We then went further and said, common KPIs, common cloud infrastructure, security architecture, you know, all these kinds of things.
PIERRE LE MANH
So people were allowed in theory to not use the platform.
STEPHAN PRETORIUS
Initially, there was nothing. This is before the platform existed. I mean, this is like 2018, 19, right? So the second, the sort of the first phase after that was to say, you have to build to a common kind of technology standard and platform. And then sort of in the years following, it went to where we are now, which is it's one platform organization, one product organization. There's only one product organization. Nothing else gets funded. As the business went from massively federated, 450 agencies to five, that also became more natural. And as we now have one production platform and business, we have one technology stack, which is WPP Open. So it's kind of, it sort of converged the same time.
PIERRE LE MANH
It's fantastic to see how IT technology drove the entire transformation.
STEPHAN PRETORIUS
I mean, it's interesting that you say that because I've been very, very thoughtful and conscious about, are we enabling the transformation or are we driving a transformation?
PIERRE LE MANH
And so what is in the end, you're driving, pretending you're enabling or-
STEPHAN PRETORIUS
What I mean is that everything we did was sort of designed to facilitate the transformation. And in some cases, it made it really obvious that the transformation was needed. And so it probably is ultimately quite symbiotic, but it is really where transformation becomes more of an art than a science, because on the one hand, you can build things that kind of codifies a company's complexity. You can establish things that actually, you know, reinforces the complexity. But if you build something that is so simple that it isn't usable by the previously complex structure, then the body rejects the organ. So you have to be extremely thoughtful. And really what we ended up when we launched Open sort of in 21, 22, was that we built an operating system that could do both things. It could handle all the complexity of our business, but it could effectively run the entire business as a unitary brand with a single go-to-market. We found a way architecturally to build a platform that could really kind of both service the past and drive to the future. That was probably, I mean, for me, that's the single biggest kind of like conceptual nut to crack, because it seems obvious when you say it now. I promise you at the time that we were thinking through it, it was way more complicated.
PIERRE LE MANH
Are you yourself a software development professional, or how did you end up being the CTO of a company like WPP?
STEPHAN PRETORIUS
I last coded when I was about 13. Confession of a pastor CTO. I studied law and I loved technology as a kid, right? And I was sort of, you know, always took things apart and whatever. So I have sort of an analytical engineering type mind. I ended up studying law, which is really a technical humanities subject, right? I mean, it's sort of somewhere between the two. And loved it. And then ended up in media and telecoms law and started working for entertainment companies or media companies. But right at the time that the internet started getting commercialized. So my very first job was to set up a digital advertising business for MultiChoice largest TV And had fly to New York, licensed DoubleClick AdExchange. I left with a bunch of CDs that they gave me and instructions for the server to buy. And I flew to South Africa's and I bought the server and I built it and I loaded the software and we started setting ads. You know, that was in 1996. And so what's interesting is like, I sort of had this period where I wasn't really working in technology. And then from 96 onwards, I've never not worked in technology because when you are plugging in the RAID configuration in the server room and you have to kind of like figure out the tables and the data outputs, you kind of learn these things through first principles. So I was very lucky. I was just at the right time in the right place. And then for the rest of my career, I've just surrounded myself with really smart technologists who I trust. And ultimately you build up a kind of a strong intuition for technological change and transformation. And so I'd say in a way, I'm sort of glad that I'm not a developer because I think if I was, I'd get too stuck in the arcane details of languages and frameworks and do things like that.
PIERRE LE MANH
That's also a good lesson, I think, for project professionals who aspire to, at one point, get to the C-suite position that they're dreaming of, right? You have to extract yourself from the technical stuff and respect it, master it, but also be able to lead. Let's go back a little bit to WPP for a second. So you have a personal experience as an entrepreneur, very hands-on, but you're also now very senior position in a very large company and very complex company. So as you were centralizing all of this, do you feel that you've lost in corporate agility at WPP?
STEPHAN PRETORIUS
I think, you know, large companies have, they can be cumbersome in terms of execution because you have more people to rally and more people to kind of to get on board. But I think we're sort of a remarkably entrepreneurial business for our size. I think it might be because we're an English company and the English for all their faults are actually quite quirky and quite sort of individualistic. And you can make decisions quickly and you can, in very small groups, quickly come to a decision and then act on it. And so we're not hierarchical. We're not very stars and bars oriented, if you know what I mean? Like many, especially American companies are very hierarchical in the sense of, you know, you don't.
PIERRE LE MANH
Not all of them.
STEPHAN PRETORIUS
No, not all of them. But, you know, sort of they are. And maybe in other cultures too, you know, where you don't criticize your VP, you don't criticize your SVP. And WPP is not like that. I mean, we're quite a sort of informal company in the sense that people show up as individuals. You can make a very big difference as an individual in the room, you know, if you know what you're talking about and have a strong point of view.
PIERRE LE MANH
So you managed to protect the psychological safety as you are centralizing at the same time?
STEPHAN PRETORIUS
A great way of putting it. And I think psychological safety is a really important part of running a creative organization. I guess all professional service organizations, but particularly a creative organization, because real creativity only thrives in safety. If people don't feel psychologically safe, they won't create. I think it's really important to be able to create that psychological safety. And it's very much something that permeates through the entire organization.
PIERRE LE MANH
Yeah, very good. I'd like to talk a little bit more about the workforce evolution. So with all this automation, the changes that AI are bringing, how do you see the future of the knowledge workforce? Will people still have jobs? Maybe we can talk about the junior people first, right? So we all learned doing basic stuff that seems to be increasingly automated. So how are we going to train younger people? Or are we overestimating or underestimating their ability to just adjust?
STEPHAN PRETORIUS
I'm very optimistic about the young generation, because I think, candidly, there are a lot of things that people our age have done throughout their careers that there's sort of no point in training people to really do, right? I mean, for sure. Everything about the amount of time that you spend kind of like, you know, sort of filling in spreadsheets and what else. There's a lot of knowledge work that is inherently inefficient, you know, a waste of time, doesn't require a lot of creativity or brain cells or thinking. I mean, it's just stuff that you have to get through, right? And so I don't actually worry about that. And I don't worry about the young generation, because what I've observed in the people that we're hiring into our company now is that they come into the workforce with a mindset of solving problems with AI tools. We run this fantastic creative technology apprenticeship program. And I'll never forget, I was sitting in one of the training sessions, and the lecture was teaching people how to do Blender scene descriptions, right? In, you know, 3D scenes in Blender. And the minute of, you know, sort of 22-year-old put a pen, she says, confidence with ChatGPT. And of course, you can. And then a whole new workflow emerges and no one's going to ever do a blend the scene description again. You're going to do it with AI. And so I think, you know, I really don't worry about the young generation. I think the nature of the work changes, but I think their agility in terms of learning the tools is just incredible. But what I also think is important is that as important as learning the new tools and new technologies are is a really strong grounding in particularly humanities. Knowledge work, everyone focuses on STEM because that's technology and development and whatever else, engineering. But actually, the humanities are really important for knowledge work. Critical thinking, psychology, you know, philosophy, theory of knowledge. I mean, incidentally, the French are actually really well positioned for this, right? Because, I mean, how do you write essays? What's the three steps of essay writing? Thesis, antithesis, synthesis, right? I mean, and if you don't write an essay like that, you fail.
PIERRE LE MANH
I go out in France, as you know, so I still write my speeches. Anything I do is in three parts. It's just natural.
STEPHAN PRETORIUS
By the way, I actually think the French are really well positioned for the AI era because of that.
PIERRE LE MANH
They think, you know, the French believe that they're very well positioned because of their engineers.
STEPHAN PRETORIUS
Yeah, I think it's the opposite, actually. You know, if you speak to someone like Arthur Mensch would actually agree, right? I mean, it's sort of a- So I think the kind of combination of being fluent in AI technologies and at the same time, having strong groundings in humanities, in creative disciplines and so on is really important because it's the combination of the two that gives you the human perspective, gives you taste, gives you judgment, gives you context. And I see that coming through in buckets, right? And curiosity. Very interesting.
PIERRE LE MANH
So that's for the young people. What about people of our generation? Are we safe because we'd be retired by the time AI procedures?
STEPHAN PRETORIUS
That's one way to look at it. No, either that or we've become true experts, right? So I think there's a sort of a category of people who throughout their careers are focused on understanding a certain domain, whether it be law, accounting, marketing strategy, et cetera, really well. And they get to a point where there are incredible domain experts. And as long as you can learn to harness that domain expertise and build AI tools, that reflect that, that encode your brain in an AI agent, for instance, I know it sounds crazy, but people are doing that, then you'll be fired as well because your expertise is unique and your understanding of the domain, your understanding of what is good input data or not, your understanding of what is good output, you know, will make you valuable. What I worry about is frankly the people in the middle.
PIERRE LE MANH
So who are they? What do they do? And what can they do to evolve?
STEPHAN PRETORIUS
Let me first say, I worry about the people in the middle because I think there's a sort of a layer where there's a kind of a tier of the workforce who've become senior because they were once good at their craft, but stopped doing their craft. And they've become managers, organizers, they've become senior, often well-paid, but frankly, they've stopped thinking. They've stopped really being curious about things. They've stopped learning. And that's the group that I worry about because I think a lot of that work and a lot of those processes can be done with AI and will be replaced by AI. So what can you do? It is a great saying that the best time to start learning about AI or the second best time to start learning about AI is today. The first best time to learn was yesterday. I think the message I give everyone is really immerse yourself in these technologies and apply your domain knowledge and your skills to it and do something, learn something every single day. It's like daily habits, take even 15 minutes every day and incrementally learn something new every single day. Because working with AI, this concept of how do you work with AI isn't like something you can just do a crash course over a weekend. It is an incremental process of micro epiphanies. And so as you're going through the process, you're working with something and then you realize something and you go, oh, wow, that's amazing. And then the next day you build on that, right? But it takes a bit of time for your brain for that to settle, to sediment, and then move on. And I'll give you a great example. So we have this fantastic tool in WPP Open. It's our creative ideation canvas. And this tool, it looks a bit like Miro in terms of the user interface, but it's a many to many multimodal, multi-agent system. So you can ask any question. So like, how do I sell more iPhones in France? Starting question. This system then starts giving you recommendations for what you can do a SWOT analysis, look at the market sizing, and you build all this out. And so it's a fantastic kind of like flow process of ideation. And so, you know, we built it to be multimodal, which means that it can generate images, can generate video, it can generate sound files, but it can also interpret all those things. And then because it's many to many, you can take any artifact on the canvas and combine it with any other artifact on the canvas and say, what would this audience description, persona description, think of this concept? And it gives you an answer. You can also say, what would this audience think of these three Instagram ads that are artifacts on the canvas? And it would interpret both the images and the text and tell you what to do. Or you can say, take this brief, take the sound file, take this sample image, and create a storyboard for a film based on all these things combined. And it interprets the style of the image, the style of the music, and the brief, and writes a storyboard.
PIERRE LE MANH
That's fantastic.
STEPHAN PRETORIUS
Now, what's wild about that is that it's effectively the multimodal intersection of different media types, an image file, a sound file, piece of text. And technically, it's just a vector embedding. So I'm going through this whole thing, building all these different flows and sort of realizing how powerful this is. And then I remembered something that one of my developers said to me in a very late night WhatsApp one day. It was like, you know, people seem to be sort of like enormously long rants about things.
PIERRE LE MANH
Really?
STEPHAN PRETORIUS
Yeah, yeah, yeah. No, completely. And I mean, sort of like good rants, you know, they sort of...
PIERRE LE MANH
Exactly. They're passionate.
STEPHAN PRETORIUS
Yeah, they're passionate. And, you know, this guy, I've been working with him for 20 years, and he concluded his rant. This was literally in 2023, beginning of 2023. He just ran by saying, Stephan, I'm now convinced in the future, most creativity will simply be a vector embedding.
PIERRE LE MANH
How does that fly?
STEPHAN PRETORIUS
Yeah, exactly. At the time, I didn't quite realize what he meant. And then when I did this multimodal combination, the many-to-many combination on our Canvas platform, I remembered that comment and I said, and I thought to myself, Lin was right. All we are doing is we are creating very smart vector embeddings between different concepts and different images in this amazing user interface. So it sometimes takes time, even if you're very close to it, even if you're building the things, even if you are the one directing the direction of it, for your mind to adjust and for those epiphanies to happen to the point where you actually feel really comfortable and confident working with AR.
PIERRE LE MANH
Stephan, thank you so much. That was a very inspiring talk for our audience.
STEPHAN PRETORIUS
Thank you, Pierre.
PIERRE LE MANH
And good lessons to learn here, like humanities, develop your Critical thinking, keep learning, be curious, encode your knowledge if you can, and we'll be fine.
STEPHAN PRETORIUS
Absolutely. It'll all be fine.
PIERRE LE MANH
Thank you.
STEPHAN PRETORIUS
Thank you.