I spent the better part of a day recently sitting in sessions about how AI is reshaping commerce, listening to platform leaders, brand-side marketers, and retail executives talk about what's working and what's about to change. The throughline was impossible to miss: shopping doesn't look the way it used to, and advertising has to catch up.
A recent BCG study found that shopping is now the third most popular reason consumers use generative AI, and that usage grew 35% between February and November 2025 alone. Consumers are using AI to brainstorm purchases, compare products side by side, and get recommendations tailored to their exact needs. They're starting searches with "ideas for," "where should I," and "which," and they expect something more useful than a list of blue links in return.
The industry has started calling these people "super-empowered consumers." I think that's directionally right, even if the label is a little buzzy. What's undeniable is that these shoppers behave differently, which has real implications for how brands invest in media, creative, and measurement.
You'd expect AI-assisted shopping to streamline the process, cutting out steps and making it more linear. But the data says otherwise. According to Google-commissioned research from Ipsos, shoppers who use AI interact with an average of 12.1 touchpoints during a purchase, compared to just 4.2 for shoppers who don't. That's nearly three times as many points of contact.
But, crucially, those extra touchpoints don't slow people down. Three in four AI shoppers say the experience helps them make faster and more confident decisions. Consumers are essentially offloading the heavy cognitive work (the product comparisons, spec research, and review scrolling) to AI, and then acting with the speed of an impulse buy backed by the confidence of a thoroughly researched one.
This creates a compressed but high-frequency window of influence for advertisers. Brands have more opportunities to show up during the consideration process than they did before, but the decision cycle is shorter. That means every touchpoint has to earn its place. Generic awareness plays won't cut it when a consumer's AI assistant is doing the vetting in real time.
In an AI-mediated purchase experience, the consumer isn't browsing a results page and choosing which link to click. They're asking a question ("what's the best running shoe for flat feet under $150") and expecting a direct, useful response. If your ad can surface at that moment with the right product, the right price, and the right information, you're completing the process rather than interrupting it.
Now think about the implications of this change in shopper behavior for product feeds and structured data. Q&A-style information that AI can pull from to answer specific questions about your brand and products is becoming a baseline requirement. If your product data is thin or poorly structured, AI systems don't have enough to work with when a consumer asks a specific question. The brands showing up in agentic commerce environments will be the ones whose catalogs are rich enough for an AI to recommend.
This extends to how brands think about discoverability broadly. That same Google study found that 70% of Google Search shopping queries now focus on discovery, exploration, or research. Keyword targeting and landing pages to capture demand are giving way to a new experience where your brand's structured data and content depth determine whether AI surfaces you at all.
One of the most interesting conversations I heard recently involved a major apparel retailer describing their move toward what they called "systems-based thinking," with flywheels that connect data, content, activation, and commerce into a single operational loop rather than managing each as a separate workstream.
Their approach reflects where agentic commerce is heading. The concept of Universal Commerce Protocol, or UCP, is designed to give AI agents a shared framework for discovering inventory, evaluating offers, and completing transactions across platforms. When that infrastructure is in place, the line between creating demand and converting it disappears. A consumer can ask an AI assistant about a product, get a personalized recommendation, and complete the purchase within the same conversation thread.
The brands preparing for this now are the ones building on a unified data foundation that informs their creative, product feeds, media activation, and measurement. They're automating the connection between a query and a conversion so that when an AI agent asks about their brand, the answer is immediate, accurate, and transactable.
This doesn't mean every brand needs to overhaul its entire tech stack tomorrow. But it does mean that the distance between your product data and your media strategy needs to shrink. If your media runs on one set of data and your product feeds run on another, you're creating gaps that AI agents will bypass entirely.
The measurement conversation is evolving just as fast. In an AI-mediated environment, clicks become less reliable as a signal. When a consumer's AI assistant evaluates five brands, recommends one, and completes the purchase inside a single thread, there's no click on your ad, no landing page visit, and no traditional attribution event. The value was created, but the measurement framework doesn't see it.
This is why incrementality testing is gaining ground over multi-touch attribution as the primary way to evaluate media performance. Incrementality answers the question that matters most — "did this investment create outcomes that wouldn't have happened otherwise?" — without relying on click paths that increasingly don't exist. Real-time signal tracking and outcome-based metrics give advertisers visibility into what's actually driving revenue, even when the path to purchase runs through an AI intermediary.
The brands getting this right are pairing incrementality with real-time audience tracking or MMMs and feeding those insights back into their activation strategies. Measurement isn't a reporting function anymore. It's an optimization input that needs to move at the speed of the media it's evaluating.
Agentic commerce can feel like a lot, with new protocols, consumer behaviors, and measurement requirements. But the first moves I'd recommend making are practical:
Audit your product feeds and structured data for conversational depth. If an AI asked a specific question about your product (i.e., shipping, materials, sizing, sustainability), would your catalog have the answer? If it doesn’t, focus first on filling those blanks.
Then connect your media strategy to your product data so that what you're promoting in paid channels matches what AI systems can surface and transact on. A disconnect here means you're paying to create demand that AI agents can't convert.
Revisit your measurement approach with an eye toward incrementality. If your primary reporting still relies on click-based attribution, start building the infrastructure for outcome-based evaluation now, before the gap between your metrics and reality gets wider.
And if you're not sure where your brand stands on AI readiness, kick that conversation off sooner rather than later.
We work with brands every day on exactly this, connecting data, media, and measurement so the system works as a whole, not as a set of disconnected parts. Reach out if you're ready to start mapping it out.