Is your brand really selectable in a world driven by AI?

Most of us spend far too long scrolling Netflix before choosing something to watch. The platform has thousands of titles, yet many of us still end up watching whatever appeared on the homepage while dinner gets cold. For me, more often than I’d like to admit, that means Dinner Dates.

That says a lot about modern choice. We feel as if we are exploring everything, but the options in front of us have already been narrowed. Netflix shows what it thinks is most likely to work, then we choose from that edited version of the world.

Commerce is moving in the same direction. For years, the Internet was built around active discovery: comparing, clicking, eventually deciding for themselves. Now AI platforms are starting to shape the shortlist before someone reaches a brand’s website, and sometimes before they have properly worked out what they want.

The new first filter
Someone who once typed a query into Google might now ask an LLM for a recommendation and accept the answer without visiting a brand’s site. AI Overviews now appear on nearly half of all Google searches globally, and organic click-through rates have fallen on queries where they appear. The answer is increasingly served before the click becomes necessary.

The next phase takes this further. Amazon’s “Buy for Me” feature lets its AI agent find products on third-party sites and complete the transaction inside the Amazon app. ChatGPT has launched instant checkout with Etsy and Shopify merchants, while Perplexity has integrated PayPal so users can buy directly from search results. These examples are still largely US-first, but they point clearly to what is coming for UK brands.

AI commerce is becoming the equivalent of the Netflix homepage. It surfaces what it can understand, trust and act on.

Consumers are still involved, of course. MTM’s recent Media360 study found that 60% of people prefer to direct the algorithm rather than let it decide entirely. Younger audiences are especially active in shaping what they see, using search behaviour and content signals to train the feed around them. The direction may come from the consumer, but AI is increasingly assembling the options. That leaves brands with a simple question: are you selectable?

The problem with separate channels
Most marketing organisations are poorly built to answer that question. Paid, organic and commerce platforming still tend to sit in separate teams, with separate targets and separate reporting. That has always created friction. In AI-mediated commerce, it creates a bigger problem because AI platforms evaluate the connected picture around a brand.

Customers experience one brand. AI platforms do the same. They are looking at whether products are structured clearly enough to understand, whether content signals genuine expertise, and whether pricing, stock and catalogue data are reliable enough to act on. If those signals are messy or inconsistent, the brand becomes harder to recommend, regardless of media spend or brand awareness.

Paid, organic and commerce need to work as one system. Paid media gives brands real-time insight into which messages resonate and which audiences are moving. Organic now has a wider role than traditional SEO, because AI platforms use indexed, structured content to answer natural language questions. Commerce infrastructure then gives those systems the confidence to act.

Making products easier for AI to understand
Amazon’s “Buy for Me” agent needs clear product identification, accurate pricing, consistent stock signals and clean categorisation before it can complete a transaction. A messy catalogue creates hesitation, and hesitation reduces the chance of being selected.

Product feeds therefore need to become richer. Many feeds are still built around a title, price, category and a few basic attributes. AI platforms need more context. A sofa described only as “two-seater, grey, fabric” tells a system what it is. A sofa described with room suitability, care needs, style comparisons and use cases gives AI much more to work with.

Product detail pages need the same treatment. People are asking tools questions such as “what is the best X for Y use case?” AI answers from content, so pages that explain trade-offs, suitability and points of difference will have an advantage over pages that simply describe features.
Structured data and taxonomy matter for the same reason. Review signals, availability data, clean categories and clear product names all help AI crawlers build confidence. Patchy attributes make it harder for systems to match a product to the right query.

The race has already started
AI is already helping consumers choose, and it will soon take on more of the buying process itself. The infrastructure is being built through AI search, conversational commerce and shopping integrations across major platforms.

If an AI platform was asked to recommend your best-seller today, would it find enough clear, trustworthy information to include it? Or would ambiguous categorisation, thin product copy and inconsistent pricing push it towards a competitor with a cleaner catalogue?

The brands getting this right now will be easier for AI systems to read and recommend as these journeys become more common. Those that leave messy data, thin content and disconnected channels in place risk being filtered out before the customer ever sees them.

Meanwhile, I’ll be on my couch watching Dinner Dates, waiting for Netflix to recommend something different. It probably won’t. It already knows what I want.

Mark Byrne is director of paid media at Brave Bison