AI is already doing your customers’ research. Soon it will do their shopping. Most marketing plans, and most agencies, are still designing campaigns for a human scrolling a feed.

Every marketing plan rests on one assumption: gain the attention of your ideal customer enough times with the right offer, at the right time, and they’ll eventually buy.

That assumption is starting to crack. AI tools have taken over much of the research, and in some markets, they’re taking over the purchase process too.

When the thing choosing between you and your competitor doesn’t watch ads or scroll Instagram, a lot of what brands and agencies have historically done to drive revenue stops making sense.

Step one: AI Is Already the First Salesperson

Someone types “I’m looking for a chart plotter for my boat” into ChatGPT. The tool asks what kind of boat, what it’s used for, and what the budget is. By the end, the buyer knows exactly which products to consider and goes straight to that brand’s site to make the purchase.

A parent needs a rugged carry case for a child’s school tablet. The AI tool takes the initial request and turns it into AI-friendly criteria: a rugged, built-in screen protector, kid-friendly, with a shoulder strap. Three options come back, and one gets ordered the same day.

No retargeting ad. No comparison shopping or email capture. The touchpoints brands have spent years paying for were skipped entirely. The case that won didn’t have the biggest budget. It had details written clearly enough in the right places for an AI agent to find and offer the best options within seconds.

Step two: AI Becomes the Buyer

Every AI agent within an LLM is a glorified salesperson. A recent Harvard Business Review article on AI agents in China explains the process in detail. People hand an agent their payment details, give it a task, and get on with their day. It orders lunch, books a flight, or finds the perfect running shoes and more.

The US isn’t there yet, but Google and Shopify have already published a shared standard that lets a purchase occur inside the AI tool, without the buyer visiting the brand’s website.

For brands and agencies, this raises a series of questions.

For brands: if a machine creates the shortlist, what are we marketing to, and how do we ensure we’re included?

For agencies: what are we now being paid to influence, and how do our strategies evolve to meet this new market need and dynamic?

Spending More vs Winning With AI

For years, paid media could cover almost any weakness. Weak rankings, thin reviews, no word of mouth? Increasing the budget could carry you until more permanent strategies were implemented to correct the gaps.

This doesn’t work very well when an AI agent does the first round of choosing. The agent compares specs, prices, reviews, and reputation across every option, and a bigger budget doesn’t change what it reads or determines is important.

There’s a risk for smaller brands too. If AI rewards whoever has the most content in the AI-prioritized places, the biggest players win on volume. The opening for niche brands is to be the clearest, most specific answer in a narrow space before the answers and the top content settle in.

Brand Still Matters

Some believe agents will shop like prioritized shopping lists based on requirements, and brand will stop mattering as much. Others argue people buy with emotion and loyalty. Someone who has always worn Nike won’t be thrilled when an unknown knockoff shows up in the recommendations just because it fit the criteria.

Both are partly right. People will still tell their agents what they like or pick the familiar name from the shortlist. The brand work just must happen before the agent is ever given a request, and it has to connect directly to what drives the sale.

The Rinse and Repeat Agency Is in Trouble

Agencies that tend to run campaigns with a similar approach across accounts are in for an awakening. Taking a client’s budget, running ads on the usual channels, optimizing to one number, sending reports and repeating, will not work with AI.

That model falls apart when decisions happen inside AI tools that never see the ad. And when an agent buys on someone’s behalf, much of what agencies report on disappears. No pixel fires and there’s no click path. Just a sale out of a black box.

The agencies that retain and increase their value will ask harder questions. Are we speaking to a person, a machine, or both? What does the AI need to know to recommend this product? What signals still exist when tracking goes dark? This thinking is closer to business consulting than media buying, and it’s what brands should be expecting and paying for.

What Brands Should Do Now

• Make sure AI can find you. Check ChatGPT shopping eligibility and your Google Merchant Center feed, whether you’re on Shopify or not.
• Write product pages for a machine that’s taking notes. Size, material, use case, who it’s for, what makes it different. If it isn’t written down, AI can’t find it or use it when comparing options.
• Keep investing in your brand. People will still tell their agents what they prefer. Make sure your name is the one they include in the upfront ask.
• Treat reputation like revenue. Ask for reviews, respond to the bad ones, and keep your eye on Reddit and forums in your category.
• Split your search strategy. Separate learning searches from buying searches and invest your money where purchases happen.

What to Ask Your Agency

• What’s our plan to win when AI does the first round of choosing? A good answer covers product data, reviews, and AI visibility, not just ad budgets.
• What happens to our reporting when tracking goes dark? Ask which signals they’ll rely on and how they’ll model what they can’t see.
• Who owns our product information? If no one on the agency side is responsible for how AI reads your product pages, that’s a gap. Accurate product pages take a joint effort.
• What have you stopped doing, and what should we be doing that we’re not? If everything looks the same as a year or two years ago, that’s your answer.

The Bottom Line

A new decision maker now sits between brands and their customers, and it doesn’t respond to what marketing has relied on for decades. The brands and agencies that win will make a brand easy to understand, trust, and already be on the customer’s mind before the agent is asked to go research.

Your customers are already asking AI what to buy. If your agency is still optimizing only for a person scrolling a feed, it’s optimizing for a customer whose behaviors are changing, and they’re not shopping that way anymore.

For the full conversations, check out our recent episodes on AI search and agentic commerce on Contrary to Popular Opinion, a podcast from Vuja Dé Digital.

If you want an outside perspective on how ready your brand is for AI-driven shopping and buying, get in touch.