Introduction

Setting the stage and audience check-in

Well, it's great to be here. What a lovely venue. And I can see a lot of smiling faces, which is fantastic.

A couple of words about myself. So my name is Anton.

So I've done about 15 years of growth consulting. So my background is marketing science, finance, and analytics.

And about a year ago, I started my own business, which was only possible because of AI.

So what essentially we do is being a bridge between marketing, finance, and finding opportunities, right, and sizing them. And everything is powered by AI.

State of play: Where AI is today

But today, I wanted to talk to you about what are the key things I hear from the industry, from people and from secondary research, and what can we learn from mistakes and success stories as well. So a little bit about what we're going to talk. So just briefly about state of play.

For many of you, it's fairly trivial. I understand a lot of you are professionals in AI space.

Could you give me a quick nod if you're AI expert in your field? Okay, I can see. I can see a few nods.

If you're a business owner. Okay, any anybody from consulting? Okay, fantastic. Hello, guys.

Right. And any students? Awesome. All right.

So we'll talk about state of play. We'll look into the future, understand where we're actually going, and then, well, maybe discuss a little bit of what can possibly go wrong and things we can do differently in our AI initiatives.

Unprecedented adoption and usage metrics

So, well, state of play. It's not a secret.

We're really, really seeing unprecedented growth of AI technology. And just look at these numbers.

At a personal level, ChatGPT weekly active users growing like absolutely crazy. So the last number, actually last night, I've heard a new number, which is 800 million active users weekly, which I need to verify.

It's apparently like a thread. It was a YouTube video. So I was like, well, I'm not going to change slides.

But anyway, you can kind of give us an idea how quickly things are growing. And in corporate world, things are growing really fast too.

So this is a McKinsey report, 78% of organizations report that they use AI. So another interesting step for you in C-suite level, about 55% use AI one way or another. So absolutely mental.

And really, well, let's take this 100 million users arbitrary sort of benchmark and try to kind of place this and compare AI with any other technology that we've seen.

And again, look, telephone, taken 75 years. Mobile, 16. Internet, seven years. Facebook, four and a half years.

TikTok, great comparison, nine months. And ChatGPT, two months.

Why this wave is different: Four accelerants

Look, what is it telling us? There are multiple factors that played a role in ChatGPT and that type of technology to drive penetration so quickly.

Number one, it's infrastructure advantages. For ChatGPT and large language models in general, infrastructure had been available for it to be adopted really quickly. Before that, we didn't need to build infrastructure for any sort of technology like that.

Of course, we're going to run out of electricity pretty soon, but smart people are already working on that, so it's all in the making. So every single time you, well, let's think about electricity. Electricity was very helpful, but it doesn't create more electricity. Whereas what's happening with LLMs is really bringing efficiency to the table, enabling faster research and being sort of creating this strong positive reinforcement loop.

Network effects. We've seen Facebook use that. So every new contributor to the system adds something to the whole value of the network. And that's also present.

And finally, immediate utility. So it's been useful from sort of day one, which is, you can't say about TikTok or Facebook, right?

And if you think about it, anything before, they've never combined all four of these elements, right? And this unique combination of these four elements, that's what's really enabled really fast growth.

Looking ahead: What’s coming next

Now, okay, we'll know until now, things are going really, really fast, but let's try to see what's happening in the future. Despite really high penetration across corporations, it's still very early days. It's very early days across various sectors.

Early days across industries

And quite recently, I did a deep dive into oil and gas extraction, especially liquid gas. And believe it or not, at every stage of supply chain, AI is already making a difference, making a change. And they're already forecast about some sort of sentient supply chain where every element of the network is interacting with each other and decisions made instantly.

So look at these very low percentages. So what I'm trying to tell you that we're actually at the very beginning of the journey and there's a lot more to come. And it's not also about penetration.

The actual technology is becoming smarter, which is fascinating. And we have experts here that are going to tell you more about the state of play and sort of what's the cutting edge of technology.

Rapid capability curves and compounding progress

But these curves, and this is software engineering, these curves are really, really steep. And same is happening in math space as well. This was commissioned by DeepMind. And to be honest, for me personally, it's quite buckling to see how steep these curves are.

Economic impact projections

And in terms of value creation, well look, from 2025 to 2030, on an annual basis, we're expecting to get 1.8 trillion. So this is annual contribution of AI to economy. And this is the most conservative number I could find, full stop.

And by the way, 1.8 trillion is sort of more than GDP of Mexico. But then when you look at other numbers, like McKinsey, for example, they're giving annual contribution 4.5 trillion, which is absolutely mental, right?

So we're really building a new world and new economy. And by the way, this new economy and this new world will be kind of very different to what we're experiencing right now.

It can't be. It can't be otherwise.

A lot of supply chains will be disrupted. A lot of sectors will be disrupted. And the relationships will also have to regrow and be different.

Can I ask you, anybody heard about new Google protocols about payment systems between AI agents? that sort of come out a week or two ago.

Yeah, I can see a few people. And that's sort of a foundational layer to build a new economy. So the key message here is that we're at the beginning of the journey and what's coming is humongous.

Right.

Risks and unintended consequences

So when we're building all this new stuff, what can possibly go wrong? Well, of course, a lot of things. And I don't want to make this like sort of doom and gloom scenario.

But there's one thing that I'm a little bit concerned. And I wanted to draw everybody's attention to this issue. which is really, really an issue.

Labor market shifts and institutional knowledge risk

And currently, well, like recently published numbers of U.S. economy specifically, we do see an increase in production. We do see more financial output with fewer employment.

And there's recent revision of employment numbers in the U.S. And we actually, the number's been over-reported for quite a bit of time. So what we see is this effect where with fewer people, we produce more.

What we see that about 16% decline in employment in AI exposed sectors, which is a lot. And that's problematic because we are really kind of risking to disrupt the information flow in institutions where people who are coming in, they're being trained, and then move up the ladder. The worst part is that the age group that's most effective is 20 to 25.

And really, if you have friends in the industry who are trying to get into the industry, that is what's really happening. And unfortunately, this very much mimics what I hear in the boardrooms from executive interviews or in private conversations from experts.

The key topic when we talk about AI is about replacement. Central question is replacement. And specifically when it comes to financial sector, when it comes to FMCG, a little bit less so, and telecom and media, quite a bit in the middle.

From replacement to empowerment

And I hear a number, it's like plans to cut 30% or 35% of employees, which is big. And what I would encourage everybody in organizations to, really try to think about what value AI can create in a different way, in a spirit of empowerment and in spirit of creation.

In innovation, for example, practice, there are these classical 70-20-10 rules, where 70% you want to be working on your main core business, 20% driving the category, thinking about category growth, and 10% doing something very different, going outside the standard thinking. And by the way, Google motto is very much this.

You probably know that Gmail, as a product, came out of one of these 10% employees doing something different from their day-to-day job. And right now, this is as important as never to really focus our energy on empowerment and creation and start thinking outside the box rather than try to fit in and replace what we already do.

Keeping in mind that economy and the supply chains will be very different in even five years' time, it is absolutely critical to find and think about our place in this new economy. And this is only possible if you understand how the future is going to unfold and have a clear idea of which part you're going to be playing in that new world.

So there's this interesting quote about generals always prepare to fight the last war. And this is so true. Like by this replacement, we're preparing for something that's actually not gonna be relevant anymore.

And we should be focusing about planning for the future and empowering our employees and creating new business models, entering new markets. And really focusing on replacement, even how we talk about it, has really direct financial impact.

A tale of two strategies: WPP vs. Publicis

And here's the interesting case study of two companies. It's WPP Group and Publix Group. So do you know what these companies are? If not, it's okay.

They're in marketing and they hold huge, huge market shares. The point I'm making here is that they're two different communication strategies.

So Mark Creed was talking about the saving and efficiency about AI. And the policy and idea that's been brought forward really drove their share price and market share and revenue down.

Whereas their prime competitor, publicists, they are thinking about it in a different way. They're thinking about augmenting and empowering people, so with very little talk about future cuts.

And you see what it does to economy. So long story short, not only it is wise to think about this new future and spend more time to finding your place in that new world, but also it has a direct commercial impact.

Right.

Doing AI differently: How to improve outcomes

So things we can do differently. Look.

Why many AI projects fail

This stat comes from MIT research. I would take this in a bit of a grain of salt, because 95%, how do you define success or failure? But this is definitely true.

From what I've seen, a lot of AI projects are failing. And there are really four buckets why they are failing.

So there is a mismatch between technology, data, and business objectives.

And we're not going to dwell too much on these points, so we have experts in all of these elements in this room.

People and culture: The heart of change

One that's particularly important, I think, that we need to raise is people and culture. So people in culture are the heart of change and the change that's what we're going through.

So look, change doesn't happen overnight. Just to be aware that this is a journey is already a big step forward. It's a journey from knowing it to really loving it and then to really living it.

where you create a reinforcement loop in your organizations from raising awareness to building affinity to the subject where we're changing, and then to empowering these people, employees or colleagues, to drive change autonomously. So they become ambassadors for these new ideas, creating new generation of change.

A combination of the four are able to unlock growth, and that's absolutely critical for any sort of AI project. And that's the type of thinking that would be very, very useful for a lot of companies to adopt.

What works: Principles for successful transformation

And finally, what did we learn from, you know, how can we succeed in this transformation and enabling AI projects with the right culture? First, yes, change is hard. No doubts about that.

But growth mindset is absolutely fundamental. And this new thinking about that we will be changing all the time is critical.

1Experimentation is key. Enabling safe environment for your teams to experiment in a safe way is fundamental.

And I see this big difference of smaller players that are growing really fast and big players where big players really discussing, well, have you done your AI safety training, whereas smaller guys sharing the tricks and tips over the coffee.

And I'm not against safety training. It's needed, right? It's about creating safe environment where people can freely experiment. That's critical.

1And finally, top down versus bottom up. So true change is only happening bottom up, but it needs to be empowered top down.

Conclusion

Grow, experiment, empower

So ultimately, what I encourage everybody to do is to encourage ourselves to grow, experiment, and empower our teams.

Well, everything is in our hands.

Thank you.