INSIGHTS

How Google Search Agents Changes Marketing

How Google Search Agents Changes Marketing

This week, Google released personal shopping assistants designed to help shoppers "find exactly what they need at exactly the right moment." Here's how that changes marketing.

Keller Maloney

Unusual - Founder

This week, Google announced the biggest update to their platform in twenty-five years. They released "Search Agents," which run in the background, watch for what a shopper might need, and surface options at moments it judges relevant. Google described it as helping shoppers "find exactly what they need at exactly the right moment."

Although Google named them "search" agents, they are better described as a personal buying assistant. This marks the true beginning of agentic commerce: a world in which brands have to market directly to AI agents, because the AI agent is the one deciding what reaches the buyer.

What Google built

Every Google user gets their own agent that persists across a user's Google properties, holds context about their preferences and recent activity, and surfaces options when it decides the moment is right. The user doesn't have to ask for anything. The agent decides when to act, what to bring forward, and how to present it.

It’s the biggest addition to a broader rollout of AI-enabled shopping tools and infrastructure. This year, ChatGPT rolled out a shopping experience with ads within their app, and Stripe has been building payment systems so AI agents can buy on a user's behalf.

Active search versus passive shopping

Search is something a person does. You sit down, you type a query, you read what comes back, you click on one of the links, you compare options, you decide. Each step requires the user to be doing the work. Search is an active discipline that Google has trained two generations of consumers to perform.

A buying agent works differently. The user isn't searching. The user is going about their day. The agent is watching, deciding what the user probably wants or needs, and bringing it forward. The buyer's job is approval. The agent does the work—the searching and the choosing—before the buyer was involved.

This is a fundamentally new, passive version of “shopping.” This predicts a future where shopping is no longer something that you do, but instead something that is done for you by a third party.

What this means for marketing

As human buyers increasingly trust AI’s recommendations, opinions, warnings, and advice, they will delegate more of their buying decisions to their AI assistants.

AI assistants are becoming a gatekeeper between a brand and its audience. Soon, brands must convince and sell to their buyers’ AI assistants as much as—if not more—their human buyers themselves.

Brands must now market to an AI audience, which has its own beliefs about your category, your offering, and your competitors.

AI as an audience for brands

Understanding the audience is the prerequisite of any effective marketing campaign. The same is true here: what does Claude believe about your category, and on what criteria does it judge the options? Which competitor does ChatGPT think is stronger than you? Where does AI’s view of you diverge from the view your customers actually hold?

AI models are superintelligent, they have already read the entire internet (during their pretraining), and they can think deeply and reason through tradeoffs. They are a highly-informed, highly-rational, and slightly-naive consumer.

AI models have already formed opinions, biases, and beliefs about every brand. You can understand these opinions and then (a) change them with honest argument and proof or (b) position your brand and offering to appeal to them.

Where this leads

AI as an audience changes the texture of brand strategy but not the framework.

Marketing has always been about understanding what an audience understands and then shaping that understanding over time. That still applies to the new, agentic audience. The difference is in the tools and methods of understanding the audience.

The AI audience is growing at different speeds in different categories. It’s growing faster in B2B SaaS and developer tools, slower in categories where buyer taste and identity carry more weight. Teams starting now have time to develop a picture of what the AI audience believes about them before that belief starts deciding deals against them.

We wrote more about this shift in AI Is A New Audience For Brands.