Agentic Commerce Is Closer Than You Think: How It’s Changing Marketing

A few months ago, if you’d told me that nearly half of American Millennials would hand an AI agent their credit card and let it buy something on their behalf, I’d have called it a stretch. Comfortably. Confidently. “We’re years away from that,” I’d have said.

New research from RTB House says otherwise, and it suggests agentic commerce is arriving faster than most marketers planned for.

The stat that stopped me scrolling

In its latest study, “Who’s Buying? Consumer Trust in the Age of Agentic AI,“ RTB House found that 42% of US Millennials are comfortable giving an AI agent a budget of up to $250 (£185) to make a purchase on their behalf, as long as a 7-day return window is guaranteed. Strip away the safety net and that number drops to 34%. Which tells you something important: consumers aren’t blindly trusting AI; they’re trusting AI with guardrails. Approval before checkout is still the single most requested safeguard globally, and Gen X and Baby Boomers remain far more cautious (29% and 22% respectively).

But let’s not undersell what this actually means. We’re not talking about smart reordering, the AI-powered “you’re probably running low on coffee pods, want me to reorder?” use case that’s existed for years. This is genuinely new. Consumers letting an AI research, compare, decide, and purchase something they’ve never bought before, based entirely on its recommendation.

That’s not automation. That’s delegation. And delegation is a trust threshold most brands haven’t had to earn before.

AI has already out-trusted your social feed

Here’s the part that should really get marketers’ attention: the same research shows that leading AI tools, such as Google AI Overviews and, in particular, ChatGPT, are now more trusted by shoppers than TikTok, Instagram, and Facebook. They’re even outperforming newspapers and established media brands for trustworthiness.

That’s a genuinely remarkable pivot. Social platforms spent over a decade building shopping trust through influencers, UGC, and native commerce features. AI chat interfaces have done it in a couple of years, largely by being useful rather than persuasive. Google AI Overviews and ChatGPT currently hold a clear lead in consumer trust; Claude and Grok are still catching up in both US and international markets, worth watching, because that gap won’t necessarily stay fixed.

The discovery layer is shifting under your feet

It gets more interesting. 59% of US consumers and 52% of non-US consumers say AI is good at surfacing brands they didn’t already know about. And the demographic pattern isn’t what you’d expect: Gen X (61%) and Baby Boomers (62%) are among the most engaged users of AI for discovery, arguably because they’re the ones most fatigued by scrolling social feeds for recommendations and happy to just ask a direct question and get a direct answer.

Put these findings together and a pattern emerges: consumers are increasingly skipping the search bar altogether. They ask ChatGPT, Gemini, or Perplexity a question, and instead of ten blue links to sift through, they get one synthesized answer. Sometimes that answer names three brands. Sometimes it names one. Either way, the AI has already done the shortlisting, and if your brand isn’t in that shortlist, you were never in the running.

This is exactly why GEO exists

This is where Generative Engine Optimization (GEO) stops being a buzzword and starts being infrastructure.

For twenty years, SEO was about ranking on a results page a human would scroll through and judge for themselves. GEO is a different game. It’s about making sure your brand, your product data, and your credibility signals are structured in a way that an LLM can confidently cite, recommend, and act on, often without a human ever seeing the ten alternatives it discarded.

That means:

• Structured, machine-readable product data. Specs, pricing, availability, and reviews need to be clean, well-structured data, consistently published, not just for Google’s crawler, but for the retrieval layer behind AI Overviews, ChatGPT browsing, and Perplexity.
• Third-party credibility signals. LLMs lean heavily on independent reviews, comparison sites, forums, and press coverage to validate claims. Owned content alone won’t get you cited.
• Clear, answer-shaped content. Write like you’re the source an AI would quote: direct claims, defensible facts, and structured comparisons, not just brand narrative.
• Consistency across the web. Conflicting information about your product across different sources erodes the confidence an AI needs to recommend you with certainty, and low-confidence answers get hedged or dropped entirely.

What this means for marketers, right now

We’re at the point where the question isn’t “should we care about AI search?” RTB House’s data makes clear that shoppers already trust these tools. More than the channels we’ve spent a decade optimizing for. The real question is: when an AI is asked for a recommendation in your category tomorrow, does it know enough about you to say your name? If the honest answer is “we’re not sure,” that’s the brief.

Agentic commerce isn’t a distant hypothetical anymore. It’s a budget line. And the brands that win in this new retail landscape will be the ones that plan for it now. 42% of Millennials have already mentally approved. The brands that show up in that $250 shortlist won’t be the ones with the loudest ad spend. They’ll be the ones the machines already trust.