Brands are constantly vying for consumer attention. But, as generative AI platforms increasingly influence decision-making, could brands shift to competing for the attention of… AI agents, instead?
Maybe! It’s a fascinating question posed by Aravind Srinivas, founder of Perplexity. Perplexity, which is widely regarded as the leading gen AI search engine (sorry, Google!), has been breaking new ground in delivering ad experiences alongside gen AI search results, launching contextual ads in November.
In a wide-ranging interview with India’s The Economic Times, Srinivas ponders how advertising might evolve in a world dominated by gen AI.
What if, he says, “each transaction itself can be monetised. Like the AI agent takes the plane ticket payment, not just a vendor. And I think that’ll lead to many interesting outcomes.”
He went on to describe a potential new model for advertising born out of this: “You could also think of advertisement models there. You could think of the vendors paying extra for giving certain special deals to the agents. One thing I’ve been just hypothetically thinking about is how it would look, if the ads are at the level of agents—the user never sees an ad. So on Google, the different merchants are not competing for users’ attention. They’re competing for the agents’ attention.”
Srinivas also added another layer of complexity, saying, “Agents can actually have advertisements thrown at them and you can teach the agent, ‘These brands are malicious. Just ignore them even though they try to advertise to you.’ And the advertisers don’t know that—that’s a confidential interaction between you and your agent.”
This “out there” concept imagines a world where brands compete for the attention of AI agents instead of users. Users could guide their agents to prioritize certain brands they trust and block others they dislike. This feedback would remain private, hidden from advertisers, hopefully creating a more personalized and less intrusive (better!) ad experience.
Why This Matters:
While this model could, in theory, better align ads with user preferences, it raises major questions about feasibility.
How would AI agents handle competing offers from brands? How would they balance user preferences with advertiser incentives? Ensuring fairness and transparency in this new paradigm would also be a massive challenge. Srinivas’ vision outlines an ambitious, if not very complex, future that could change the relationship between brands, users, and AI.
Although this vision might feel more like a thought experiment than an imminent reality, it deserves credit for rethinking how ads could be delivered. With nearly half (45%) of the U.S. population already using generative AI—a number that will only grow—traditional ad models may not translate seamlessly into AI-driven platforms.
Experts React:
AdTech X was a bit skeptical of Srinivas’ potential model but it sparked some interesting thoughts. Here are some of the more best takes from Eric Seufert, Rob Leathern, Amit Goel, Ari Paparo, and Lou Paskalis:
Our Take:
Again, Srinivas gets points for thinking differently. For now, adapting existing ad formats to fit generative AI might work, but as the technology evolves, more fundamental changes will likely be needed to make sure ad experiences align with how people, you know, actually use AI.
Some version of a brand → agent → consumer model seems inevitable. At least we think so. But whether it will take the form Srinivas describes, who knows. He is, however, becoming one of the more interesting voices in AI-based advertising. Plus, Perplexity ships quickly, so we’ll see where this goes in 2025.
