Hey, welcome back to the C Space Studio. I'm James Coteki here at CES 2026, and I'm here with Zeta Global's co-founder, chairman, and CEO, David Steinberg. David, welcome to the C Space Studio.
Thank you for having me, James.
Okay, so define Zeta Global for us for a second. What brings you to CES?
Well, first of all, we're at CES for the same reason a lot of people are. Everybody's here.
That's exactly right. But Zeta Global is a big data and artificial intelligence platform company, and our software helps very large enterprises more cost-efficiently run their marketing and CRM.
Yeah. And so tell me about what that looks like in practice. Give me some examples of how that might show up.
We help very large enterprises to better manage their marketing. So Forrester did a study recently that for every dollar that an enterprise spends on our software, we return them between six and $7 in revenue.
So 600 to 700% return on investment by using our software and our data.
So does that imply that there's a lot of inefficiency that's happening before you show up? Tell me about where some of the biggest inefficiencies are that you're addressing.
Yeah, by the way, I wish I could say it's just us being exceptional. That's part of it, but a lot of it is. And if you think to sort of the crux of artificial intelligence, a lot of people are talking about AI. At Zeta, we started programming an artificial intelligence in 2017.
So what we're really able to do is target an individual who's not just actively in market for our clients' products, but we can put other pivots on top of it. By way of example, this person might be very interested in your product, but they can't afford it or they won't get credit approved for it.
We're also able, when you look at a lot of marketing today, a lot of it is just retargeting on the internet. So you click on something, you buy it, and they follow you around for two weeks with an ad. The crazy thing is a lot of that is massive inefficiency because you've already bought the product.
Because we plug into our client CRM systems, we know you just bought. So we remove you from the marketing to save our clients money. Those are just some examples.
And then how does that show up for the way that your clients are using it? Is it an LLM? Is it a report? Is it plugged into the CRM that they're already getting? Is it writing along? How do they actually experience it on their screen?
Yes. Okay. We have one user interface that allows a client to do three main use cases: customer acquisition, customer retention, customer monetization, and all of the activation for any digital methodology, whether it's programmatic, video, connected TV, mobile, social, all runs through our platform. And then there's one reporting interface that gives you attribution for every single spot you ran, all the way from the beginning of the journey to the purchase of the product.
You said that you've been doing AI for almost 10 years now. Are there things that you've learned about AI that other people haven't caught up to yet? Are there things that you know about it that people right now are just now finding out?
Yes, by a lot. And obviously I'm CEO of the company, I have to say that. But the truth is, if you think about the future of AI, and I think people forget this, a hundred years ago there were about a thousand companies in America manufacturing cars.
Today there are five. You're going to see a very similar thing in large language models. They're going to become very ubiquitous. There'll be a few major winners, and there'll be a lot of people who do not win. At Zeta, what we have found is that the data that you feed into the algorithms is the biggest differentiator. And what we're doing is we're taking our client's first-party data, we're merging it with our first-party data, and the algorithms are learning off that in a very private environment that never has any data exhaust. So we're actually on our fourth version of our AI agent studio. Our clients have built thousands of custom AI agents on top of our platform by using our agent studio. Most of our competitors are on V1.
Okay. Tell me about what the definition of agent is in this case, because my impression is people use the term to mean different things, and different levels of autonomy are implied by that depending on who's saying it.
Totally. I mean, now some people are calling bots agents, right? I mean, if you think about it, to us an agent is performing a singular task as a component of a supply chain or a process chain. In this case, let's talk about marketing, right? So we might have one agent that will literally look at every customer you've created in the last 30 days, look at the 240 million Americans in our data cloud, and say, 'This 2 million people are doing that exact same thing.'
Okay. Another agent might say, 'Here's the best place to target them.' The coolest thing about our platform is we've strung three full agents together. The third agent in real time is informing the first two with an attribution capability.
Okay. So they're all getting smarter, and we just announced that we're rolling out Athena, which is our voice-enabled conversational super agent that'll control the entire platform. And yesterday after the market closed, we announced that we've done a full partnership with OpenAI, and we're going to be integrating Athena as native to their application layer.
