Welcome to day three of Innotribe, and we've got a fantastic day once again for you. And this session is going to be awesome. In our first session in this double take on the future of value, we examined the technological landscape and how intelligent contracts could work in a new era of digital autonomous finance. In this follow-up session, we'll look at how the first wave of digital assets has already created programmable money, and that second wave, agentic AI, will create intelligent, self-directing finance.
So the session will explore a future where AI agents — systems capable of setting goals, planning actions, and autonomously executing complex financial transactions — take control and manage trillions in assets. Once we move beyond using AI as an assistant to having AI act as an operator or an agent, what happens when a proprietary AI is given a bank account and the ability to lend, to trade, and to manage compliance all on its own? What about those profound challenges of accountability, ethical bias, and explainability as core financial functions are delegated to code that adapts and learns in real time?
So I'd like you to join me in welcoming to the stage Azor Baros, managing partner at Parallax Ventures; Martin Müller, head of artificial intelligence and agentic AI for financial services EMA from Microsoft — that was a big one — Sergey Nazarov, co-founder of Chainlink, and our moderator, Leah Solomon, executive producer for FF News and CNBC. Please give them a warm welcome.
Thank you, Inis. And good morning, ladies and gentlemen. It's fantastic to be back. As mentioned, this time yesterday we had part one of Future of Value, where we looked at the role of intelligent contracts. And today, for part two, we are going to explore what happens when money becomes intelligent. And I have a fantastic panel. As mentioned, I will just reiterate who we have here on the stage: Sergey Nazarov, co-founder of Chainlink. Welcome. Welcome to the three of you. Thank you for being here today.
Sergey, I wanted to start with you because you have already been on one of these Future of Value panels. It was about three years ago. How have things evolved in this agentic AI space since then?
I mean, three years ago, I don't think there was agentic AI. So it's evolved quite a bit and quite rapidly. I think the past panel I was on was about digital assets and blockchains. I think that's evolved quite a bit. In that last panel, we talked about, you know, whether blockchains and digital assets would be widely adopted. And I think this year and even last year at Sibos, you know, Swift has a ledger, DTCC has an app chain, TCH has a ledger, so on and so on, right? We can just go on: who has a blockchain. So that was the topic of the previous panel and what we discussed. And I think a lot of what was said there has happened: that digital assets and chains have come to be a widely accepted technology that FMIs and banks all need to implement for themselves.
Sergey, Martin mentioned their digital assets. Tell us your view on how agentic AI is transforming that side of things.
Yes. So I think there's two versions of the world there. One version is AI agents as an interface to digital assets. So that means the user, or, you know, the institutional middle office or back office people instruct the AI agent to take action, and then the AI agent takes action itself on the digital asset, right? And that's the infrastructure we provide: is the ability for the AI agent to interface with the tokenized equity, the stablecoin, the tokenized deposit, move it around for collateral management. You know, the more AI agents are directed to do things, the more they're going to need infrastructure through which they can do those operations for the tokenized digital asset side. That's one version of the world.
There's another version of the world where — and we're thinking a lot about this internally — what if AI agents become their own user group, right? So what if there are just AI agents out there who make a lot of decisions with minimal human input? You know, some amount of human input, but a smaller amount than we expect or that we're used to right now. That's a user group of its own, right? And then you need to make digital assets attractive for them. You need to make stablecoin and tokenized deposit payments attractive compared to traditional payments, and you need to provide infrastructure and access points for them to use that set of assets, right? The digital assets. That's an interesting possible world. It's not something that's clearly happening right now, but at the rate at which AIs are evolving and all that's happening, and the amount of decision-making that's being ceded to them, you could be in a world where you have AIs that are just making a huge amount of purchasing and asset allocation and asset movement decisions with relatively minimal human input. So I think those two worlds.
And Sergey, would you agree with that timeline? And are you equally as optimistic?
I mean, I'm freaking out about the example where there's $450 running shoes. I was just sitting here like freaking out about that. I'm like, 'Wow, that's a lot.' You know, I am wearing running shoes and these things cost like a hundred something. I get them on sale. Hell, I buy like eight pairs. I'm sorry. After you said $450 for... I think I got the value in euros and I just put dollars in. I don't know what these running shoes are. I hope that doesn't happen because I really don't want to pay that much for running shoes. I'm still stuck on that.
Everything from there just kind of like... But I think, you know, people go down the path of least resistance, and there's optimization of all these UIs, and everyone's like, 'Let's fix the loading speed of our website and let's make our website easier to use.' And they do these screenshots of like, you know, there's this famous screenshot they did of Craigslist or of like a bank website, and they point out this thing's getting replaced by this startup, this thing's getting replaced by that startup on your website, right, on your bank website. And what are those websites that are replacing a piece of the bank website? They're just optimized, vertically focused companies focused on that part of the bank's business, right? That was like the previous theory. And why would that work? That would work because those startups would make a better user experience, right? They would make a better user experience for that one part of the bank's business than, you know, that bank could do for all parts of their business. And basically, with AI, I haven't seen a screenshot like this, but now the screenshot is like the whole website. It's just like one arrow: AI.
So yeah, I think that's true, right? Because who wants to do extra work? Like the scenario they both described: get me the $450, unbelievably US dollar inflated, you know, the world's on fire version. I, you know, so you triggered me with this $415 sneaker, or the Ralph — get me the Ralph Lauren thing. Thank God you didn't say that was expensive. But, you know, yeah, I think everyone's just going to say, 'Do the thing, AI thing. Do the thing.' And so that AI is just going to make a lot of purchasing decisions. And whoever optimizes their offering for that AI to make the purchasing decision in their favor is going to win. And whoever is sitting there playing the old game of like, 'Let me make my website better for people to click buttons. Let's get those humans to click those buttons' — like, okay, go ahead. I mean, I don't think it'll work, but I don't think humans are going to be clicking buttons anymore. I guess what I think — that's what we're all saying, right? The button-clicking humans are now going to be like, I see this thing saying everything I say in real time. Is that... oh yes, that's freaking me out too, right up there. Yeah, like, so you know, that's it. This thing knows exactly what am I saying. It knows everything. I know everything. Yes, we're testing it in real time and it's working very well.
