The AI Purchase Funnel Just Closed. ChatGPT, Gemini, and Meta Now All Sit Between Your Brand and Your Buyers.
Meta is testing an AI shopping tool today to rival ChatGPT and Gemini. That completes the trifecta. Three platforms covering most of the world's AI-active users can now research, compare, and recommend products on behalf of buyers — and none of them route through your ad budget.
Today, Bloomberg reported that Meta is testing an AI shopping research tool for US users, positioned directly against similar features already live in ChatGPT and Gemini.
This is not a feature story. It's a completion.
ChatGPT launched shopping research in November 2025. Walmart integrated directly. Target launched a conversational shopping experience inside ChatGPT. PayPal partnered with OpenAI to power checkout inside the chat window. Google had already woven shopping into Gemini's interface. And now Meta — 3.35 billion daily active people as of Q4 2025 — is testing its version.
This isn't three companies building the same feature. This is the AI purchase funnel closing.
The Frame Everyone Is Missing
Most marketing teams are reading this as: Meta launches shopping, interesting, file under platforms to monitor. ChatGPT shopping is a ChatGPT thing. Gemini shopping is a Google thing. Separate experiments, separate responses.
That framing treats these as parallel. They are not. They are three entry points to the same shift: the first moment in a buying journey — the one where a person asks "what should I get?" — has moved off the open web and into AI chat interfaces.
When your buyer opens ChatGPT and types "what's the best [category]," they get a recommendation. Not ten blue links. A recommendation. Same for Gemini. Same, within weeks, for Meta AI.
None of those recommendations come from your Google Ads account. None come from your retargeting budget. They come from what the AI learned from sources it trusts.
What AI Shopping Actually Runs On
Here's where most people get it backwards: AI shopping agents are not searching the web in real time to figure out which brand to name. They are drawing on a model trained on the same corpus every AI system uses — news coverage, editorial reviews, analyst reports, the body of credible written record about your category.
Walmart partnered with OpenAI to make its products findable inside ChatGPT. That matters for product discovery in consumer retail. But if you're a B2B brand, a SaaS company, a professional services firm — anything without a SKU — the shopping AI learns what it knows about your category from the same source it always did: the written record in publications it treats as authoritative.
A company with consistent coverage in Forbes, TechCrunch, and relevant vertical trade press surfaces in an AI shopping recommendation. A company with no earned media presence, regardless of ad spend, does not.
I wrote about this mechanism when AI agents started handling B2B vendor research: these systems don't discover vendors the way a human does, by browsing and comparing. They recall what was written about them in publications the model treats as credible. The AI shopping race just made that mechanism universal — not just a B2B phenomenon.
What Most Brands Are Missing
Most are either ignoring this entirely, or reacting to each platform announcement as if it's a separate tactical problem.
It's one problem. The AI layer that sits before the first human click now covers the most widely used chat interfaces on earth. The signal that determines whether your brand appears in that layer is the same whether the platform is ChatGPT, Gemini, or Meta AI.
Third-party editorial credibility. Not paid placement. Not SEO copy. The publications these AI systems were trained on — which happen to be the same ones that have shaped human brand perception for decades: Forbes, TechCrunch, Reuters, WSJ, and hundreds of vertical outlets that AI engines treat as authoritative.
When Meta's AI shopping tool goes live and someone asks it to compare vendors in your space, what does it know about your brand? If the answer is "not much," that is not a Meta problem. It is a visibility gap that opened the moment ChatGPT launched shopping in November. Meta's announcement today just makes it impossible to rationalize ignoring.
What Changes From Here
The purchase funnel used to start with a search. Then with social. Now, for a growing share of buyers — both consumer and enterprise — it starts with a chat. Research tracking enterprise buyer behavior put 47% of enterprise buyers starting vendor research with AI ahead of Google. Consumer behavior is tracking the same direction.
That number will keep climbing as Meta AI rolls shopping out to its user base.
The brands already building for this understand something the laggards don't: AI shopping agents are not a distribution channel you advertise into. They are a retrieval layer. You earn placement in them the same way you always earned credibility with informed buyers — through third-party media that the AI learned to trust.
That mechanism is Machine Relations. PR's original operating principle, now applied to machine readers. The publications have not changed. The audience reading them has.
Three platforms. The funnel is closed. Where is your brand when someone asks?
Related Reading
- AI Visibility for SaaS Companies: How to Get Cited by ChatGPT and Perplexity
- Machine Relations for Fintech Companies: How to Get Cited by ChatGPT, Perplexity, and Financial AI Engines
The visibility audit runs your brand against the citation signals these AI platforms actually use. Five minutes to find out where you stand.