From chat to checkout: what you need to know about UCP

The advertising industry hasn’t changed all that much over the past 10–20 years. Actually, it never really changed drastically, until now. Say hello to UCP (Universal Commerce Protocol).

Sure, we’ve seen new formats and placements come and go — Reels, Performance Max, and various flavors of automation — but structurally, things have remained fairly stable. Not that long ago, paid social and paid search were separate worlds, run by different specialists. Setting up and auditing accounts took a smaller team and was fairly complicated and time-consuming. (Raise your hand if you felt like you decoded humanity when setting up conversion goals in your Google account in 2005.)

What’s happened over the last year, though, feels different. More has shifted in digital advertising recently than in much of its history.

That’s why something that happened this weekend is worth paying attention to — not because it was loudly announced or packaged as a major launch, but because it quietly changed one of the assumptions e-commerce has been built on for years.

Until now, AI chats have mainly been places where people ask questions, compare options, or — let’s be honest — even ask for medical advice. Chats have been useful early in the buying journey, acting as a kind of referral or shortcut to information. The actual purchase, however, still happened somewhere else — on a product page, inside a checkout flow, through ads, or via other paths most of us take for granted.

That’s now starting to change.

As of January 11, Google made it possible to complete a purchase directly inside Gemini. Instead of clicking through to a store or buying from sponsored products, the transaction happens inside the chat itself.

Through a partnership with Walmart, Gemini users can now browse inventory, receive recommendations based on previous purchases, and check out without ever visiting Walmart’s website. You can ask for a specific product or a broad category, see what’s available at that moment, get suggestions for related items, and complete the purchase in one continuous flow.

However, when the conversation itself begins to function as the storefront, the shift is greater than it seems. Product discovery, comparison, and decision-making begin earlier in the journey, often before a brand has a chance to present itself in its own environment.

This isn’t just Google and Walmart

The bigger story isn’t the partnership. It’s the infrastructure underneath it. Google and a group of partner companies are launching a new initiative called the Universal Commerce Protocol (UCP). Think of it as a shared language that enables AI agents to communicate directly with retailers, payment providers, and commerce platforms.

And it’s not a small club. Target. Shopify. Etsy. Wayfair. Best Buy. Home Depot. Macy’s. Visa. Mastercard. Amex.
More than 20 major players agreed on a common commerce standard before this thing even properly launched.

Here is a good time to grasp.

If UCP takes off, it means any AI assistant can discover products, compare prices, personalize recommendations, and complete purchases — without sending users to the store, whether it’s on Shopify, a website, or elsewhere. Your website stops being the default point for checkout.

“But people won’t actually buy inside chat.”

That was the safe assumption. And it held — right up until it didn’t.

During the 2025 holiday season, AI assistants influenced roughly 20% of retail sales, representing more than $260 billion in revenue. Not just “helped people research.” Actually influenced purchasing decisions.

What’s even more interesting: traffic coming from AI assistants converts dramatically better than most channels. When users click through from tools like ChatGPT or Perplexity, conversion rates are reported to be several times higher than those from social traffic.

Why?

Because the decision already happened. By the time someone clicks, they’re not exploring. They’re confirming. If that sounds familiar, it should. It’s branded search energy — at scale.

Distribution beats features (again)

Here’s where Google’s move starts to make uncomfortable sense. Gemini didn’t win because it’s the “best” AI. It won because it’s everywhere. It’s baked into Android. Embedded in Search. Living inside Gmail. It shows up where people already are — not where they have to go. 

That distribution advantage is starting to show. Gemini’s market share has grown rapidly over the past year, while competitors that rely on standalone apps are feeling real pressure.

For commerce, that matters more than model quality. 

You don’t need the smartest assistant. You need the one people already use.

The gap between demos and real shopping

Here’s a telling data point.

Despite massive usage, only a small fraction of AI conversations today involve products that can actually be purchased. That gap — between curiosity and commerce — is exactly what UCP is designed to close.

And analysts are paying attention.

Estimates vary, but most agree that 10–20% of e-commerce transactions could be handled by AI agents within the next few years. Some forecasts go much higher.

Even the conservative viewpoint suggests a significant shift in how buying decisions are made — and where they occur.

Amazon saw this coming (of course)

Amazon’s response was predictable and very on-brand. They exited Google Shopping entirely. Instead of participating in an open protocol like UCP, Amazon is betting on a closed ecosystem: their own marketplace, their own assistant, their own rules. It’s a classic platform strategy. Own the customer. Own the data. Own the transaction.

But it comes with risk.

If UCP becomes the standard and most major retailers adopt it, AI-mediated shopping could start excluding Amazon by default. That’s a bold bet — even for them.

What this means for Shopify brands and performance marketers

This is where things get practical. If AI assistants become the first — or only — interface shoppers use, a few things change fast:

  • Discovery moves upstream
  • Fewer comparison shoppers reach your site
  • Creative becomes structured data
  • Price, availability, and trust signals matter more than persuasion
  • Attribution gets even messier

For Shopify brands, this cuts both ways.
Open protocols mean you’re not forced into a single marketplace tax. You can show up where intent lives without giving everything to Amazon. But it also means you’re no longer just optimizing for people. You’re optimizing for systems that decide on people’s behalf.

That’s a different game.

This isn’t about replacing stores

Your site still matters. But it may no longer be where decisions are made. Those decisions are increasingly happening upstream — inside interfaces designed to reduce friction, not showcase brands.

The funnel didn’t disappear.
It moved.

And the marketers who adapt won’t look radical. They’ll just look obvious in hindsight.
Because when chat becomes the storefront, commerce doesn’t get louder. It gets decided earlier.

In this article

Related content

In marketing, understanding the concept of market saturation is pivotal for strategizing and achieving sustained business growth. This guide will This guide will delve into the nuances of market saturation and its implications for marketing strategies.
Learn why maximize conversions can outperform target CPA and ROAS for better Google Ads results.
Performance Max, a revolutionary campaign type in Google Ads designed to maximize your advertising performance using Google’s advanced machine learning technology. This guide dives into Performance Max, its unique features, and how it can help your digital marketing efforts.

Stay updated

Get on the (email) list!

    
       

Stay updated

Join our newsletter