ChatGPT and the Changing World of Ecommerce Discovery

We're thrilled to be joined on the blog this week by friend of the agency Sam Wright, founder of Blink SEO. Sam is one of the industry's leading voices and shares his valuable insight into the future of the online shopping ecosystem.

In April, OpenAI made public its plans to integrate shopping results directly into ChatGPT.

This didn’t come out of nowhere - it's been obvious that AI tools would double down on product discovery for some time. But this announcement marked a clear shift.

Then, in May, Google’s I/O keynote doubled down on it. Shopping results featured prominently throughout, with examples showing how users can browse, compare, and explore products directly in the interface. SGE is no longer experimental – it’s getting a full rollout, with over a billion users soon to be exposed to AI-powered product discovery directly in search.

Discovery via ChatGPT is no longer hypothetical. It’s happening. At Blink, we’re already seeing traffic and conversions across our projects. And as usage grows, this is likely to become one of the more important new channels in eCommerce - especially for product discovery.

A New Landscape

What’s emerging is a more conversational, intent-led way of shopping. Someone might ask ChatGPT, “I need trail shoes under £100 for wide feet - ideally something sustainable.” From there, they can browse, compare, and even check out - all without visiting your site.

That’s a real change in behaviour. Discovery is moving further up the funnel, and the path from awareness to conversion is flattening. In some cases, it’s just a few prompts long. For brands, the practical implication is simple: you’ll need to be visible in these new environments. And visibility will depend on how well your product data matches user intent.

When someone types a question into ChatGPT, it doesn’t look at your website the same way a search engine does. If your products aren’t described in a way that makes sense to the model - or the data doesn’t exist in a structured format - they’re invisible.

So while your site still matters, your product feed is increasingly what represents you. What used to be backend detail - categories, attributes, tags - is now central to how you’re discovered. Brands need to start thinking of their product data not just as an internal resource, but as the outward-facing expression of their catalogue.

The Shift is Already Happening

This isn’t a theoretical shift. Traffic is already showing up. We are consistently seeing traffic and sales from chat.openai.com, and we expect to see a lot more of it in the months ahead.

That makes this a practical challenge, not a strategic “trend to watch.” It’s already affecting visibility, and soon it’ll affect performance. If your product feed is patchy, flat, or overly generic, you’ll be less likely to show up in AI-led environments - whether that’s ChatGPT, Google’s Search Generative Experience, or whatever comes next.

Product Data Becomes a Shared Priority

The challenge is that most organisations aren’t set up to manage this. Taxonomy and tagging often sit somewhere between product, merchandising, and development - but they’re rarely owned properly by any of them. And marketing teams, who are usually responsible for driving visibility, aren’t always close to the product data.

As discovery becomes more decentralised, this disconnect becomes a risk. If you want to show up in channels powered by AI, someone has to take responsibility for making sure your data is structured, complete, and kept up to date.

That means agreeing who owns it. In many cases, it’ll need to sit closer to marketing - but supported by product and dev. The best outcomes tend to happen when cross-functional teams agree on a structure, define the core attributes, and put processes in place to maintain them over time.

From Dev Task to Growth Lever

It’s worth saying clearly: this is no longer just a dev task. A few years ago, you could treat taxonomy as backend hygiene - something you fixed when it broke. That’s no longer true. Product data now sits upstream of multiple growth channels: SEO, paid shopping, internal site search, and now AI discovery. It needs to be accurate, flexible, and expressive - not just technically correct.

That means treating it with the same attention you’d give a landing page or ad campaign. You’re shaping how your brand appears in new discovery environments - and in some cases, whether it appears at all.

What to Do Next

For most brands, the right first step is to get a clearer picture of where things stand today. That means:

● Reviewing your taxonomy. Are your categories clear and mutually exclusive? Do they reflect how customers shop?

● Auditing your attributes. Do you consistently capture materials, use cases, fit, target audience, values, and other key filters?

● Cleaning up tags. Are they being used in a structured way, or just accumulated over time?

● Assigning ownership. Who’s responsible for maintaining this - and do they have the access and authority to make changes?

Once that’s done, you can start building a more structured approach: one that supports not just your website, but every other environment your products might appear in - including AI tools.

What this all means

AI-led discovery is going to change how people find and buy products. It’s already happening. For eCommerce teams, the challenge now is operational: how do you make sure your catalogue is structured in a way that supports this shift?

The answer isn’t complicated - but it does take work. It means putting proper ownership in place, treating product data as a growth lever, and making sure your catalogue is something machines can understand. Brands that get ahead of this will see the upside first. The rest will play catch-up.

Written by Sam Wright, founder of Blink SEO — a Shopify marketing agency known for its industry-leading SEO work with large catalogue stores.