AI assistants like ChatGPT are changing how people discover, research and compare products. For retailers, this raises an important question: How do you influence a customer journey that increasingly happens outside the channels you control?
The purchasing decision is moving upstream
Imagine someone looking for a wool sweater for a weekend at a Norwegian cabin in January.
Traditionally, they might have searched Google, visited a few online stores, compared products and eventually made a decision. Today, they can ask ChatGPT which materials are suitable, what to look for in a good sweater and which brands offer the best options.
By the time they visit an online store, they may already have a shortlist.
And this isn't just a theoretical shift. According to Statistics Norway, 69% of Norwegians aged 16–79 used generative AI within the previous three months in 2026, up from 54% the year before.
Shopping behaviour is changing too. A 2026 Prisjakt survey found that 32% of Norwegian consumers had already used AI tools while shopping.
That doesn't mean AI is about to replace traditional e-commerce. But it does suggest that brands need to pay attention to what happens before customers arrive.
Being discoverable is not the same as being useful
It's tempting to see this as another visibility challenge. Just as brands have invested in SEO to appear in search results, they now need to understand how to appear in AI-generated recommendations.
That's certainly part of it. But I think there's a more interesting opportunity.
ChatGPT can already recommend products using publicly available information and merchant product data. Brands don't necessarily need a dedicated integration to appear in those recommendations.
The question is what happens when a customer needs something more specific.
Can the assistant recommend the right size based on a brand's actual sizing guidelines? Check whether a particular colour is in stock? Suggest compatible products or find availability in a nearby store?
This is where the distinction between simply being visible and being genuinely useful becomes important.
Your competitive advantage is what AI doesn't already know
General-purpose AI assistants are increasingly capable of helping people compare products. Simply giving them another product description doesn't create much additional value.
But your business has information and expertise that an AI assistant may not have access to.
Your sizing rules. Your product compatibility data. Your live stock levels. Your knowledge of how customers actually use your products.
Through integrations such as ChatGPT apps built using MCP (Model Context Protocol), brands can make selected capabilities available directly within an AI conversation.
For example, instead of simply recommending a wool sweater, an assistant connected to a retailer's systems could help a customer choose the appropriate size, check availability and continue towards a purchase.
That's a much more compelling proposition than building an AI integration for its own sake.
And it creates opportunities beyond product discovery, including customer service, personalisation and connecting digital experiences with physical stores.
Better data creates better AI experiences
There's a catch, of course.
If your product information is incomplete, your inventory data is unreliable or your systems don't communicate properly, connecting them to an AI assistant isn't going to magically solve those problems.
If anything, it risks exposing them in a new channel.
This is why I believe the conversation about AI commerce needs to extend beyond the latest tools and integrations.
Many retailers are still working to improve the data foundations that support SEO, conversion optimisation, personalisation and performance marketing. Those same foundations will be essential for AI-assisted commerce.
Better AI experiences start with better information, not necessarily more AI.
The good news is that improving these foundations creates value across your existing channels, regardless of how quickly AI shopping develops.
Start with a problem worth solving, not a technology worth testing
The opportunity is exciting, but the best place to start isn't necessarily a large technology investment. It's finding one valuable use case and testing what becomes possible when your own product expertise meets AI.
A sensible first experiment might involve just one AI assistant, one product category and one capability that goes beyond what a generic AI response can provide.
Take sizing as an example. Could an assistant use your actual product data and sizing rules to guide customers towards a better choice?
If you want to explore that opportunity, I'd start with three practical questions:
- Is your product data accessible and reliable? Can your systems provide accurate product information, pricing and availability?
- Can the experience lead somewhere useful? Is there a practical route from recommendation to product selection and purchase?
- How will you know whether it works? Can you measure usage, customer engagement and ultimately commercial outcomes?
Where the necessary APIs and data are already available, a focused prototype can potentially be developed in a matter of weeks. Our approach is to start with a controlled test before considering a wider launch.
The objective shouldn't simply be to prove that an integration works. It should be to learn whether it solves a meaningful customer problem and creates enough value to justify further investment.
The opportunity is bigger than ChatGPT
It's easy to get caught up in the individual features being released by OpenAI and other AI providers. And given how quickly these platforms are evolving, today's capabilities may look quite different six months from now.
But the underlying commercial shift is more important than any particular feature.
Customers are finding new ways to research products, evaluate alternatives and decide who to buy from.
For brands, that means thinking beyond how to attract visitors to an online store and towards how to make their expertise, products and services useful wherever those decisions happen.
The opportunity isn't simply to create another route to your website. It's to become a more valuable part of the customer's decision-making process.
And that seems like a much better place to start.
Learn more about our approach to commerce, or get in touch to explore what a practical first experiment could look like.
