Shopify’s leadership has been telling investors something that caught my attention: AI search is driving more traffic and sales to their merchants, not replacing Google. As someone who builds bots for a living, my first reaction was skepticism. But after sitting with the data they’ve shared, I think this has real implications for how we architect conversational AI and shopping experiences.
What Shopify Actually Reported
The numbers are straightforward. AI-driven sessions to Shopify stores tripled year-over-year in Q2 2026. Referral sessions from AI chatbots — clicks coming from ChatGPT, Perplexity, Google Gemini, Microsoft Copilot, Claude, Grok, and similar tools — grew more than 8x year-over-year on the platform. And here’s the part that matters most for merchants: AI-referred shoppers convert better and spend more.
Yet organic search remains dominant. Google isn’t losing its throne. What’s happening instead is that AI tools are creating an additive channel, a new stream of qualified buyers who arrive with higher intent and deeper context about what they want.
Why This Matters If You Build Bots
I spend my days building conversational agents and smart bots for businesses. So when I see data like this, I’m not thinking about it as a headline. I’m thinking about architecture decisions, product recommendations, and what this means for the bots we ship to clients.
If AI-referred traffic converts better, that tells me something specific: users who interact with an AI assistant before landing on a product page have already done their research. They’ve asked questions, narrowed their options, and arrived ready to buy. The AI did the top-of-funnel work that traditionally required blog posts, comparison pages, and retargeting ads.
For bot builders, this flips our optimization priorities. Instead of building bots that try to close sales inside the conversation itself, we should be thinking about bots that prepare buyers and then hand them off cleanly to a merchant’s store. The conversion happens downstream, but the AI session is where the decision gets made.
Structured Data Is Your Bot’s Best Friend
One practical takeaway I keep coming back to: if AI tools are sending qualified traffic to e-commerce stores, then those stores need to be legible to AI systems. This means:
- Clean, structured product data that AI crawlers can parse
- Schema markup that gives conversational AI enough context to recommend products accurately
- API-friendly product catalogs that third-party bots can query
- Consistent naming conventions and attribute tagging
If you’re building bots that surface product recommendations — whether that’s a custom shopping assistant or a plugin for a larger AI platform — you need solid data pipelines feeding your agent. Messy product feeds produce messy recommendations, and messy recommendations don’t convert.
Google Isn’t Going Anywhere, But the Funnel Is Changing Shape
The narrative that AI search would cannibalize Google traffic has been loud for the past two years. Publishers have felt real pain from AI-generated summaries eating into their click-through rates. But e-commerce appears to work differently. People still search Google for products. They also now ask AI assistants for recommendations. These behaviors coexist.
What I find interesting from a bot-building perspective is that this creates a multi-entry funnel. A shopper might discover a product through a Perplexity query, validate it through a Google search, and then return via a ChatGPT conversation a day later to actually purchase. Attribution gets messy, but the signal is clear: AI is a net-new discovery channel for commerce.
What I’m Building Differently Now
This data has changed how I approach two types of projects at my shop:
First, for clients who sell physical products, I’m prioritizing bots that optimize for referral quality over in-chat conversion. The goal is to match user intent with the right product, provide enough context to build confidence, and then send them to the store with a clear path to purchase.
Second, I’m spending more time on the data layer. Making product information accessible, well-structured, and easy for external AI systems to consume is becoming as important as the bot logic itself. If ChatGPT or Gemini can’t accurately represent your product, you’re invisible in this new channel.
The Takeaway for Fellow Builders
Shopify’s data confirms something I suspected but couldn’t prove with my own client metrics alone: AI assistants are becoming a legitimate acquisition channel for commerce. Not a replacement for search. An addition to it. For those of us building smart bots, that means our work isn’t just about customer support automation anymore. We’re building the front door to a purchase decision. That’s a responsibility worth taking seriously.
🕒 Published: