AI Discovery: What E-Commerce Founders Need to Understand

The ground is trembling in e-commerce as a new way for customers to discover products is emerging at speed, and unlike most "new" channels, this one is fundamentally different from everything that came before it.

When someone asks ChatGPT for a product recommendation, or asks Claude what the best option is for their specific situation, they're engaging in a discovery mechanism that doesn't exist anywhere else. It's not algorithmic like Meta's feed, it's not search-driven like Google, it's conversational, AI-powered, and genuinely high-intent.

For advertising, this channel is still in beta. Access is inconsistent and geographically limited (the UK experience is particularly patchy at the moment). But the direction is clear, and the opportunity for early movers is real. This article breaks down what's actually happening, why it matters, and what founders should be thinking about now.

WHAT AI DISCOVERY ACTUALLY IS

When a customer asks an AI tool for a product recommendation, they're providing that AI with context about what they need. The AI then recommends specific products, often with links to purchase options. If your product gets recommended, you get traffic and potential sales from that recommendation.

This is different from ads in a meaningful way. It's not an interruption or a promoted placement. It's a genuine recommendation from a tool the customer trusts.

OpenAI is currently testing product feeds in ChatGPT's Ads Manager. Brands can submit their product information, and when users ask ChatGPT for recommendations in relevant categories, those products may be suggested. Other AI tools are exploring similar integrations, but we have less information on these at the moment.

The real opportunity isn't just about ads within existing AI tools, it's about becoming discoverable when customers use AI for problem-solving, which is increasingly how people research purchases.

WHY THIS IS DIFFERENT

Every other discovery channel optimises for one of two things: engagement (like Meta) or intent capture (like Google). 

AI discovery is different because it's conversational. A customer isn't scrolling a feed or searching keywords. They're having a conversation with a tool, explaining their specific situation, and asking for a recommendation. The context is richer, the intent is clearer and the moment of recommendation feels more personal.

There's also no saturation yet. Meta and Google have billions of advertisers competing for attention. Their capabilities are immense and sophisticated, but the auction is fierce. With AI discovery still in beta and at an early stage, there's no auction yet, no rising costs and no algorithm fatigue. Early movers have genuine room to test at lower cost.

THE CURRENT STATE (AND WHY THE UK IS PATCHY)

OpenAI launched product feeds in ChatGPT's Ads Manager in the US, and brands are testing it. They submit product information, and when relevant queries come in, their products get recommended.

In the UK, access is inconsistent and limited. Some brands have access, but many don't. The rollout is happening, but slowly. This is partly intentional (OpenAI is testing carefully) and partly because the UK regulatory environment and market dynamics are different.

However, the pattern is clear, this is rolling out geographically and it will reach the UK more comprehensively. The question isn't whether it arrives, it's when, and whether you'll be ready.

WHY THIS MATTERS FOR E-COMMERCE FOUNDERS

Customer intent in AI discovery is extremely high. Someone isn't passively browsing, they're actively asking for help to solve a problem or find a product. They're in a decision-making mindset, which is high-intent traffic at its purest.

There's also a contextual advantage that's different from other channels. When someone asks an AI tool for a recommendation, they're in problem-solving mode, actively seeking help. A product recommendation that fits their need feels like a natural solution to their question rather than an advertisement interrupting their day. That shifts how people receive and consider it.

For founders, this means you're not just buying impressions or capturing intent at the moment someone searches, you're getting recommended at the moment someone is asking for help. Those are different moments, and they convert differently.

WHAT FOUNDERS SHOULD DO NOW

If you can't access the beta yet, don't wait passively. Start preparing now for when you can.

First, audit your product data. AI recommendation systems rely on rich, accurate product information. Your product descriptions, specifications, reviews, pricing and images. All of this needs to be clean and detailed. An AI needs to understand not just what your product is, but what problems it solves and who it's for.

Second, think strategically about which of your products make sense for AI discovery. Not every product is a good fit. Products that solve specific problems, that have clear use cases, that benefit from personalised recommendations, these are the candidates. A generic commodity product might not work as well as a niche solution to a specific problem.

Third, monitor the beta and watch how the platforms evolve. If and when you get access, understand how it works before committing budget. Test with a small feed, learn what recommendations you're getting, understand the customer journey and measure the conversion and value.

Finally, build the infrastructure to manage product data across channels (your Shopify store, Meta catalogue, Google Shopping feed etc). As more platforms integrate AI recommendation, you'll need clean, consistent product information feeding multiple systems.

WHAT’S CHANGING

We're moving from AI incidentally recommending products when people ask, to AI recommendation becoming a deliberate, trackable, optimisable commerce channel. At the moment, if someone asks ChatGPT for help finding a product, it might recommend options, but there's no way for you as a brand to track that, measure it, or optimise for it. There's no infrastructure connecting your products to those recommendations systematically.

What's emerging is the infrastructure to make this intentional. Product feeds submitted to AI platforms, Ads managers where you can see what's happening, attribution tracking so you understand which recommendations converted, and the ability to optimise your product data specifically for how AI systems evaluate and recommend products.

For founders, this means your products could eventually be discovered at moments you're not currently competing in. Someone possibly isn’t using ChatGPT because they want to shop, they're using it to solve a problem. Your product could get recommended as part of that conversation and that's a different discovery mechanism than anything available now.

TIMELINE

Now, in summer 2026, AI discovery is in beta. Geographic rollout is inconsistent and most brands aren't active yet. We believe most founders see it on the horizon, but aren’t seriously thinking about it yet. If you are, you’re already ahead.

Over the next 6-12 months, OpenAI will likely expand ChatGPT access geographically (hopefully including the UK more fully) with solid measurement, clear attribution, and an understanding of customer intent. We'll also get a better sense of whether this channel actually converts and at what scale, based on early adopters' results.

By 2027-2028, if the trajectory continues, AI discovery could be meaningful for some verticals. Not dominant like Meta or Google, but a relevant part of your customer acquisition mix.

THE REALISTIC VIEW

This isn't a channel to abandon your core platforms for, Meta and Google are highly likely to remain your foundation for the foreseeable future, but this is a channel worth understanding, preparing for, and testing when you can.

The opportunity seems obvious - a new discovery mechanism with low saturation and high intent. Early movers will learn the channel before it becomes more mainstream. They'll build presence before costs rise and they'll understand customer behaviour before competitors flood in.

But it's also very early. Most brands shouldn't reallocate significant budget here yet. What you should do is pay attention, prepare your product data, and test thoughtfully when access opens up.

AI discovery is the most genuinely new e-commerce channel we've seen in years. It's not a repackaging of existing platforms, nor is it an extension of something familiar. It's a different way for customers to find products, based on conversation and problem-solving rather than algorithm or search.

You can't predict exactly how it will evolve, but you can stay alert and you can prepare. You can test when the opportunity is available and you can position yourself to move quickly when this channel matures.

We’re into the beta ads manager and beginning to build campaigns. If you’d like to discuss OpenAI ads with us, and whether you think they’re worth investigating with your business, book in a call below.