When AI sells to AI, brands win on data and identity

Your AI agent books the hotel. It recommends the insurance plan. It renews the subscription — or cancels it — based on what it’s learned about you. You barely looked up from your coffee.

Meanwhile, a brand’s marketing team is still obsessing over click-through rates on a banner ad you’ll never see. That mismatch is the story nobody in the marketing world wants to say out loud.

A sharp piece of analysis published last week in Fortune — written by Jarrod Martin of Acxiom — names it directly: when agentic AI intermediaries handle purchasing decisions, the entire game of brand influence shifts. It no longer runs through human attention. It runs through data quality, identity resolution, and trust signals that AI agents can actually parse. Strip away the buzzwords, and this announcement says one thing: if your AI agent can’t find clean information about a brand, that brand doesn’t exist in your world.

What the Fortune Analysis Actually Says

The core argument is this: the traditional marketing funnel — awareness, consideration, decision — used to play out over days or weeks across dozens of touchpoints. A consumer might see a TV ad, Google the brand, read three reviews, ask a friend, and finally buy. AI collapses all of that into a single conversation.

According to Omnicom’s Future of Brand Influence report, 45% of consumers already say AI-generated recommendations matter more than advertising in shaping how they perceive brands. And 70% believe they can become an expert in any product category just by using AI — no research phase required.

The implications for brands are serious. AI assistants are already booking hotel rooms, scheduling medical appointments, and renewing or canceling subscriptions on behalf of users — based entirely on what those agents have learned about individual preferences. Brands must now convince the AI intermediary first, before they can reach the human behind it. As Martin puts it: AI must persuade AI.

Why This Changes How Your Personal AI Agent Shops

Here’s the part that’s easy to miss if you’re reading this as a consumer rather than a brand owner: you are already on the other side of this equation.

If you’re using a personal AI agent — one that manages your calendar, handles emails, or acts on your behalf — that agent is making evaluative decisions right now. When it recommends a vendor, suggests a service, or auto-renews something on your behalf, it’s doing so based on the data it can access. Structured data. Machine-readable pricing. Clean product information. Not a beautiful landing page with hero copy and customer testimonials.

Research from House of Martech confirms what practitioners are already seeing in practice: AI agents evaluate vendors by parsing structured data, querying APIs, and comparing machine-readable pricing — without ever reading a human-written sales page. The brands that win your agent’s shortlist are the ones that prepared for a machine reader, not a human one.

The Invisible Brand Problem No One’s Fixing Fast Enough

Zoom out. This isn’t just a brand strategy story. It’s a signal about what the agent ecosystem is becoming.

Pernod Ricard’s head of digital discovered that two-thirds of Gen Z consumers were already using AI technology to research products — as of early 2026. So they audited how the major AI models actually represented their brands. What they found was jarring: one AI model categorized Ballantine’s, an affordable mass-market Scotch, as a prestige product. Years of deliberate market positioning, overridden by an algorithm working from incomplete or conflicting data.

This is the invisible brand problem. It’s not that brands are being ignored by AI. It’s that they’re being misrepresented — and they have no way to correct it in real time. If your personal AI agent asks for a recommendation in a product category and the underlying model has bad data about a brand, the recommendation it gives you is wrong. Not maliciously. Just structurally.

And agent-to-agent commerce is already happening. Experiments at Ditto show AI agents discovering, evaluating, and recommending or purchasing from other AI agents with no human salesperson in the loop. A consumer’s AI agent can communicate directly with a brand’s AI agent to discover products, negotiate terms, and execute purchases with minimal human involvement. The infrastructure exists. The question is which brands are ready for it.

We’ve been tracking the companies building in this space — the ones laying the foundation for what agentic commerce actually requires. The agentic AI companies building the future in 2026 piece we published earlier this year covers the infrastructure layer in detail. The Fortune analysis gives us the business case for why it matters.

What to Do About It This Week

When AI talks to AI, your brand’s data and identity are the message. Make sure they say the right things.

Most of this falls on brands. But if you run a business, use AI agents in your workflow, or are building agent-powered products, there are practical steps worth taking now.

• Audit your machine-readable presence. Ask an AI assistant — Claude, ChatGPT, Gemini — how it describes your brand or products. The answer tells you what data it’s working from. If it’s wrong, you have a data problem, not a marketing problem.
• Structure your product data for agents, not humans. If your pricing, specs, and positioning live only in beautifully designed web pages with no underlying structure, AI agents can’t parse them reliably. APIs, structured data markup (Schema.org), and machine-readable product feeds matter more than they ever did.
• Treat consent and data hygiene as infrastructure. The Fortune analysis emphasizes four pillars: trustworthy data, data hygiene, identity resolution, and privacy/consent. If your customer data is scattered, outdated, or missing consent signals, AI-powered interactions will degrade or fail.
• Watch for identity fragmentation. When a customer moves from mobile to desktop to in-store, most brands lose the thread. The brands that can follow that journey with consistent, permissioned data will appear in agentic shortlists. The ones that can’t, won’t.
• If you’re evaluating AI agent platforms for your own use: look for platforms that handle identity and memory consistently across sessions. The same principle that applies to brands — unified, clean, persistent data — applies to your personal agent setup.

What the AI-to-AI Era Means for Anyone Running an Agent

• AI agents are already acting as autonomous buyers — booking, renewing, recommending, and canceling on behalf of their users. This is happening now, not in 2028.
• 45% of consumers say AI-generated recommendations already outweigh traditional advertising in shaping brand perception, according to Omnicom’s Future of Brand Influence report.
• Gartner predicts 90% of B2B purchases will route through AI agents by 2028, representing more than $15 trillion in automated exchange — making machine-readable brand data a competitive necessity.
• Brands that fail to structure their data for AI parsing risk being misrepresented — or excluded entirely — from the shortlists your agent builds on your behalf.
• The winning move isn’t a bigger ad budget. It’s cleaner data, stronger identity resolution, and explicit consent signals that let AI systems represent a brand accurately and consistently.

The companies that get this right in the next 12 months will be nearly impossible to displace from AI-generated shortlists. The ones that don’t will keep paying for attention that agents are routing around entirely. That compounding advantage doesn’t wait — and the gap between ready and unready is widening every week.