Google's Shopping Agents: A Wake-Up Call for Retailers
After 20 years at the intersection of digital experience, customer loyalty and commerce data, I know one thing: the customer relationship is everything. It's why I joined Inrupt six months ago. I saw what was coming for retailers in the age of AI, and it worried me.
Last week, Google made that future arrive a little faster.
Their announcement of new AI shopping agents for retailers sounds helpful on the surface. Google is offering to embed Gemini-powered shopping assistants directly into retailer websites, promising to help customers find products faster and complete purchases more efficiently. They're positioning it as a win-win: retailers get cutting-edge AI capabilities without building them in-house, and consumers get a better shopping experience.
But let's be clear about what's really happening here.
The New Middleman, Same as the Old Middleman
Google already controls the front door to online commerce through search. Now they want to control the entire shopping journey—from discovery through purchase—by inserting their AI agent between retailers and their customers.
Think about the data flows this creates. Every product preference, every abandoned cart, every price sensitivity signal, every purchasing pattern—all of it flowing through Google's infrastructure. The company that already knows what you search for will now know exactly what you buy, how you decide, and what makes you convert.
This isn't just data collection. This is Google positioning itself as the essential intermediary in every online transaction, with visibility into customer behavior that individual retailers could never match on their own.
The search monopoly was powerful. This could be something else entirely.
The Customer Loyalty Problem
In my years in retail and payments technology, I watched loyalty programs evolve from punch cards to sophisticated data platforms. The entire point was to deepen the relationship between brand and customer. To understand their preferences. To earn their trust. To make shopping with you feel personal and effortless.
Now imagine your customer learns to rely on Gemini to do their shopping. They describe what they need, and Google's agent finds it—maybe on your site, maybe on a competitor's, maybe on Amazon. The agent remembers their preferences, anticipates their needs, handles the tedious parts of online shopping.
Who has the relationship now?
The brutal truth is that if Google's AI becomes the primary interface between consumers and commerce, retailers become interchangeable inventory suppliers. Your carefully cultivated brand differentiation, your customer insights, your loyalty programs—they all matter less when customers are outsourcing their shopping decisions to an AI that treats you as just another option in a vast catalog.
The Alternative Path
This doesn't have to be inevitable.
The businesses that will thrive in the age of AI shopping aren't the ones that outsource the customer experience to platform companies. They're the ones that develop their own capability to deliver personalized, AI-powered experiences—while maintaining direct relationships with their customers.
I'm not suggesting every retailer needs to build their own large language model. That's neither practical nor necessary. But retailers do need infrastructure that allows them to:
- Store and access rich customer data with proper consent
- Connect that data to AI systems in a privacy-preserving way
- Deliver personalized experiences across their own touchpoints
- Maintain control and visibility over how customer data is used
This is fundamentally about data architecture, customer experience and trust. The companies winning in AI aren't necessarily the ones with the best models—they're the ones who can safely feed those models the right context about individual customers. But it has to be done in a way the customer appreciates, instead of fears and resents.
What Retailers Should Be Asking
If you're a retail executive considering Google's offer—or similar proposals from other platform companies—here are the questions you should be asking:
Who owns the customer data? Not in a legal sense, but practically. Who can use it? Who can see it? Who benefits from the insights it generates? Who decides the answers to these questions over time?
What happens to customer loyalty? If customers begin to rely on a third-party AI agent, what happens to your direct relationship? How do you maintain brand differentiation?
What are the long-term dependencies? If Google's agent becomes integral to your customer experience, how much leverage does that create? What happens to pricing? What happens if they change the terms?
How does this position you for the future? Five years from now, do you want to be a merchant with deep customer relationships and sophisticated personalization capabilities? Or do you want to be an inventory supplier hoping a platform algorithm sends customers your way?
The Choice Ahead
Google's move should be a clarifying moment for the retail industry and digital businesses more generally. The age of AI intermediaries is here. Businesses have to decide whether they will build their own AI capabilities on infrastructure they control, or whether they'll rent those capabilities from platform companies who will, in the process, capture the most valuable asset in retail: the customer relationship.
I joined Inrupt because I believe there's a better path forward—one where businesses can deliver cutting-edge AI experiences while maintaining direct, trusted relationships with customers. One where customer data is treated as an asset that flows between the customer and the businesses they choose to work with, not as fuel for platform monopolies.
The technology to do this exists. The business case is clear. The standards are emerging. What's needed now is the will to prioritize long-term customer relationships over short-term technological convenience.
This moment will define which retailers still have direct customer relationships in 2030, and which have become fulfillment centers for AI shopping agents.
Choose carefully.
