Every major shift in commerce starts the same way. A new technology promises to make things more open, more efficient, more democratic, and more aligned with what consumers actually want. Then scale kicks in, incentives harden and slowly, the walls go back up.
We’ve seen it with search, social and marketplaces. Now we’re seeing it with large language models.
Recent announcements, like ads entering ChatGPT and new commerce protocols emerging point to a future in which AI systems increasingly sit between brands and customers at the most critical moment: when intent is formed.
LLMs may be becoming the next walled garden for commerce and that makes owning your customer data more important than ever.
AI is a huge step forward for commerce
There’s no question that AI and LLMs represent a meaningful step forward for consumers and businesses alike. It’s making shopping easier to navigate, helping people make sense of overwhelming choice, and reducing friction across the entire buying journey. In many cases, it’s improving the experience in ways that simply weren’t possible before.
This is a good thing. Better tools lead to better decisions, and better decisions create better outcomes for both customers and brands.
But whenever a new interface becomes the primary way people discover products and make decisions, control shifts toward whoever owns that interface. That dynamic has played out repeatedly in commerce, and agentic commerce is likely no exception.
Discovery is moving upstream again
For years, brands have adapted to wherever discovery lived whether in the Instagram feed or on Main Street, but the underlying goal has always been the same: show up where customers start looking.
Today, that starting point is increasingly conversational and agent-driven. Consumers are asking AI systems what to buy, which brands to trust, and how to compare options long before they ever visit a website or open an app. By the time they reach a storefront, many of the most important decisions have already been made.
That shift matters because discovery isn’t just about visibility. It’s about understanding. The earliest questions customers ask—what they care about, what they’re unsure of, what alternatives they’re weighing—are the richest signals a brand can have. Increasingly, those signals are being captured, but not always returned to the brands that need them most.
Why owning data matters more now
Klaviyo was founded on the belief that brands should own their customer data. Ownership is what allows businesses to learn over time, adapt as channels change, and build relationships that compound rather than reset with each new discovery channel.
AI doesn’t change that premise. It reinforces it.
The real risk for brands isn’t losing traffic or even losing channels. It’s losing context. When AI mediates customer interactions without returning insight, brands see outcomes without understanding. You might know what sold, but not why. Over time, that gap erodes differentiation and turns brands into interchangeable options inside someone else’s interface.
Owning your customer data doesn’t mean rejecting AI or trying to build everything yourself. In fact, the opposite is true. LLMs are most powerful when they can work across rich, connected systems rather than operating in isolation.
The opportunity isn’t to replace AI-driven discovery or decision-making. It’s to ensure the signals those systems generate don’t disappear into silos. When AI can both draw from and contribute to a unified understanding of the customer, it becomes a force multiplier instead of a black box.
This is why the future of commerce won’t be defined by a single model or interface. It will be defined by how well data moves between them, and––just as it always has––how deeply brands understand the customers behind the data.
Related content
