“I’m looking to buy a new jacket for the winter. I live in the UK, so it is rare that we have very cold weather, however it does rain quite frequently. I have a jacket for general use, so this one would be more for professional use: going to meetings, train travel, that kind of thing. Can you make some recommendations?“
This is the reality of shopping and discovery now. Whether you think the AI tools like Gemini, ChatGPT or similar will win the hearts and minds of people to always start their journey with them, this is the new paradigm search: even on your own website. The news and media are all over the agentic purchase protocols: ACP and UCP, where agents can take customers to a completed order. They are missing the first need of agentic success: content.
Let’s not forget my jacket query isn’t a one shot request, either. When (in this case ChatGPT) responds, there’s valuable advice and: “If you tell me (1) your budget range, (2) whether you run hot or cold, and (3) whether you’ll wear it over a suit/blazer most days, I’ll narrow this to 3 “buy-it-now” options in the right tier and style“. The conversation continues. Refines. Iterates.
Contrast this to even the best product detail pages or catalog information. Yes, we have the basics down: materials, sizes, colours. On some sites we have some great, crafted, technical features sections. Sometimes these sections explain why the feature is good (but even this is rare). To help AI answer product discovery journeys, we need a significant uplift, likely in the region of 3-5x written content.
Brands need to be so much more descriptive about the content of their products. This will ensure that when used by AI tools, or on their website, the tools can do a much better job of matching to the user’s intent and their brand showing up where it should.
Jackets are on my mind, and not only because it is -3 degrees celsius in Montreal, where I am right now. Brands like North Face, Patagonia, Carhartt, Barbour and more are all competing in this middle space I queried. Let’s see what they offer:
Many of them talk about being warm or having their own warmth ratings. Rarely will they identify a temperature range, not meeting AI where it can easily compare and contrast
Wind follows that pattern set in warmth: not specific about the kind of wind chill it is designed to withstand
Rarely is there a good Q&A section with customer submitted questions and answers (which can clarify product features / content)
There is a lot of use of proprietary labels for their technology that don’t mean anything to AI (or customers, a lot of the time)
Product descriptions that tend to be aspirational or lifestyle focused, rather than valuable or driving insight or understanding
Descriptions that are not very long: yes, customers do not read, however AI tools really do, it is the essence of how they judge
There’s a lot of work to do here. Product feed standards for AI have already changed. They are demanding more detail and depth than the previous attribute driven, sparse feeds. They are looking for rich content about products that many brands are unprepared for.
In the next post, I’ll share the even more powerful sibling of content. Context takes our ability to serve the customer to a place where we start to match user intent head on. If product content is the yin, context is the yang.