LLMs have changed how brands and products are discovered. But earning trust takes more than a recommendation.
As AI becomes a mainstream shopping interface, a new commercial reality is emerging. Consumers are using LLMs like Gemini, ChatGPT and Claude to shortlist, but before they buy, they go looking for the human proof that AI can't provide - across search, social, YouTube, Reddit and beyond.
We call this the Validation Economy - the critical space between an AI recommendation and the moment someone buys, where trust is won or lost.
So what does it take to win? Croud's latest Consumer Index, based on a survey of 4,000 UK and US shoppers, unpacks exactly where trust comes from, what consumers are looking for when they validate, and how brands can build the infrastructure to turn AI discovery into real revenue.
Discover some of the key insights from the UK survey below, and download the full report for more.
A sneak peek into some of the key findings from the latest Croud Consumer Index. Download the full report for more.
The winners won't be the easiest to find. They'll be the brands consumers believe.
Discovery now starts as a conversation, not a keyword — and it isn't a youth behaviour. LLMs build the shortlist before a preference exists.
Be describable. LLMs surface the brands they can describe precisely, so having structured, consistent content across all channels is non-negotiable.
Every recommendation triggers a hunt for human proof. The sale is now won or lost here, not at discovery.
Where consumers go for human proof after an AI recommendation.
Be verifiable. Proof can't live in one place — brief validation content as a connected system across creator, retailer and peer channels.
Shoppers are more willing to hand over the purchase than most assume. Most AI commerce sits under £100, and trust raises the ceiling.
Be delegatable. Build the right trust conditions now, across returns, data security and more, to unlock the agentic commerce opportunity.
Using Croud's proprietary BrandCI and SEOCI, we reset how Thorne shows up across AI and traditional search — moving from keyword-first planning to intent and entity-led prioritisation.
A brand tied to a need or an identity is one an LLM can describe precisely. A brand that only ever competed on price has nothing to attach to.
Leaving the AI conversation, people don't want more product detail — they want proof another human checked it. The closest to converting check the hardest.
Equity without evidence gets you recommended and then abandoned at the checking stage. Evidence without equity makes you one of several credible options with nothing making you the obvious one.
