“Time” running ads for agents; Mobian’s CEO tells how it works
AI agents are becoming a new audience for publishers. Unlike human visitors, they don’t see the ads on a webpage. The Mobian saw this as a new advertising opportunity and recently partnered with Time magazine to test the premise – a way for publishers to monetize digital news and archived content in the age of Google Zero.
Goodhart has a track record in the space. Previously, he served as CEO and Co-founder of Moat, an ad-measurement company acquired by Oracle for a reported $850 million.
I talked with Goodhart about how the new ad tech, called Agent Ads, solves a new problem and how the economics can make sense even for smaller specialty publishers. The interview has been edited for brevity.
Alisa Cromer: We are here with Jonah Goodhart, CEO of The Mobian, an exciting new ad tech company. Jonah, I’m interested in learning more about how this new technology works.
Jonah Goodhart: At Moat, our previous company, we used to say, “Bots don’t buy things.” Our technology would detect a computer that was automatically crawling a website, the original concept of a bot. Another example is Google, which actually operates the biggest bot: Googlebot.
The big deal today is that you choose your favorite AI tool, and this agent goes out and does things on your behalf.
Today, about a billion people a week use ChatGPT, and that use looks very different. It’s conversational. An average consumer is saying things like, “I want to plan a trip. I want to bake something.” AI agents are going out on your behalf, gathering information and, in some cases, making purchases on your behalf.
The world that we now live in is split between humans and AI agents.
Cromer: A lot of specialty publishers have been optimizing for search for years. If they’re B2B, they may have a lot of direct traffic and don’t really rely on search. But some of them have big archives. They’re refocusing on owned audiences, structured data, relationships, and events. How does your product apply in that universe?
Goodhart: In both of the examples —an archive or direct traffic—you’re talking about humans. What we’re talking about is actually not humans. Cloudflare announced recently that more than half of the web is not human.
How are non-humans accessing these publishers?
In that scenario, ChatGPT shows up to a publisher as an AI agent to consume content. So in that world the question is simply: How does a publisher monetize that and get paid for the valuable content they’re creating?
Cromer: You have a lot of clients—Fox, Hearst, Time. What are they actually selling?
Goodhart: Let’s take a webpage on Time, an article about the Time 100 Creators 2026. There’s a video playing. It’s engaging. You can see that it’s presented by Ally, which is a bank, with several ads mixed into the page.
The web page works for a human. No breaking news here.
But when ChatGPT or an AI agent shows up, they don’t see this page. What an AI agent sees is the code that powers the page—the underlying code.
Prior to Mobian, an AI agent would show up, take this code, figure out what it wanted from it, and throw a bunch of it out.
For example, AI agents throw out JavaScript, so they’re not actually seeing the video.
What we did in our partnership with Time was say, “What if we delivered a totally different experience for an AI agent?”
Let’s give them something that’s machine-readable. Let’s make it easier for them.
What we created with Time is what we think of as agent experiences, with an Agent Page and Agent Ads.
The page is the overall page, exactly the same, except optimized for the agent, giving them just the information they want in an efficient way. That’s number one.
We’re including all of the same metadata that the agent gets on the normal HTML page—all of the information it got before, just more efficiently and cleanly.
But we’re also delivering something new that they weren’t getting before.
If I go to the human version of the page, there are a number of different places throughout the page where you see “Presented by Ally,” with different ads.
We decided that instead of that experience for an AI agent, let’s give them one single ad unit: the Mobian Agent Ad, labeled sponsored content and supplied in partnership with the brand, the source of this material.
It is an ad unit that doesn’t look anything like what a human sees. It is a set of facts about the brand, not a slogan. It’s not marketing language.
It’s literally saying: What is Ally?
Part of the ad itself explains what Ally is: a digital financial services company offering online-only banking, including checking and savings through Ally.com and its mobile apps.
The ad unit itself goes on to provide a bunch of facts that are cited and verifiable from the brand. Ally offers access to 75,000-plus fee-free ATMs. That’s a fact according to Ally. They don’t charge overdraft fees. That’s a fact according to Ally.
The agent-optimized experience is a page with brand facts inserted that we’re calling an ad, paid for by a brand.
The big idea here, and why your audience may be interested in this, is that the number of agents going to their pages is going up. That seems to be a trend that’s continuing.
Publishers already monetize the human part of it. We’re offering them the ability to monetize the agent part of it.
Cromer: Who creates the ad?
Goodhart: We give them a tool that creates templates and drafts of that ad, but it’s ultimately the brand’s ad.
The brand ultimately owns the information. Mobian auto-generates a draft they can use once they’ve looked at it and approved it. It’s their facts.
They can change it, or choose not to use the tool and build their own.
Cromer: So publishers could sell ads against their archives using this?
Goodhart: Publishers could sell ads against their archives using this.
I’ll tell you something that, in my mind, is fascinating. Usage by AI agents, in many cases, is way higher than archive usage by humans.
Occasionally, humans will find their way into some piece of content from 10 or 20 years ago, but it’s not high volume.
We’re seeing something different with AI agents: They see a lot of value in archived content.
I think it’s because they’re getting it from a credible publisher that they trust.
The shape of the traffic, if you will, is a little bit different for AI agents than it is for humans.
Cromer: Let’s jump right into metrics and ROI. Our audience ranges from companies that have $1 million in sales up to $10 million or $15 million. Some of them are working with very small teams because these niches are very small.
It might be Parking Today—people who own and manage parking lots—or city magazines. Just about every little niche you can imagine has a publication.
Is there a revenue level that is too small for this to be profitable for them?
Goodhart: We have a different model for smaller publishers, and we’re having lots and lots of conversations with them. If we approve them and they obviously want to work with us, we’re offering a revenue-share model where they basically make incremental dollars on the inventory they make available to AI agents.
The upside for a publisher is that anything they make is extra revenue. The only downside is basically the time involved in setting this up, which is pretty straightforward. It’s their facts. The brands like it because they only pay when the Agents download their ads.
We think the value proposition here is really good for a publisher because today they don’t monetize this inventory. When they serve ads on this inventory, those ads are thrown out.
For the last couple of years is that AI models have crawled the web, taken content, and that’s it. The publisher got nothing.
Cromer: One last question: Where did the name Mobian come from?
Goodhart: The name came from Mobian Nights, a book written by my father, who was a retired English professor. In the book, he talked about context. He didn’t invent this concept, but he has a phrase in his book that says, “We cannot sidestep context.”
I thought that was a really powerful idea.
He goes on to talk about the Möbius strip, a mathematical, topological figure. One of the things my dad talks about in his book is that sometimes things that we see as different turn out to be the same thing.
He gives the example of the Möbius strip, a mathematical figure that looks like a circle but has a twist in it. Unlike a circle, on the Möbius strip you get to any other place on the strip without crossing a dimension.
Even though it visually looks like it’s two-sided, the Möbius strip is actually one-sided. That’s his example of difference being sameness.
Cromer: Thank you so much.
Goodhart: Thank you so much for having us.
For more information on The Mobian’s publisher program, contact Jennifer Bas, VP Of Operations, at jennifer.bas@themobian.com.
