Your buyers are asking Google's AI which brands they trust. The answer was written in Forbes three years ago.
Google's AI now recommends specific brands to every free US user based on their behavioral history. The brands it knows aren't configurable — they were built in publications your buyers trusted years before this feature launched.
Yesterday, Google quietly handed 75 million free US users something that changes how brand discovery works.
Personal Intelligence — previously locked behind a $20/month subscription — is now live in AI Mode, the Gemini app, and Gemini in Chrome for anyone with a Google account. It connects Gmail, Google Photos, and behavioral history to AI search, so when a buyer asks which brand of software, service, or solution fits their needs, Google's AI doesn't start from scratch. It starts from everything that person has ever read, purchased, or engaged with.
Google VP of Product Robby Stein said it plainly in the March 17 announcement: "Whether you're looking for a specific brand of sneakers you previously purchased, or planning a family getaway based on your hotel confirmations and past travel memories, Personal Intelligence helps you find exactly what you need without having to give all the context."
The coverage treated it as a personalization story. Everyone is now asking how to optimize for it.
They're asking the wrong question.
The recommendation was already made
Here's what the SEO coverage missed: Google's AI doesn't decide which brands a buyer should prefer. It recalls which brands that buyer has already encountered in sources they trusted.
The "preferred brands" in Personal Intelligence aren't surfaced from behavioral targeting in the Facebook sense. They're constructed from the editorial record — the publications that shaped what this person read, the media sources their inbox pulled from, the industry coverage that made certain company names register as credible. Before Google can recall a brand preference, that preference had to exist somewhere in the signal layer. And that signal layer is primarily editorial.
Ahrefs put numbers to this. Their study of 75,000 brands found that brand web mentions — earned editorial coverage — correlate 3x more strongly with AI visibility than backlinks (0.664 vs 0.218). The two-part analysis found that brands in the top 25% for web mentions earned 10x more AI citations than the next quartile. Brands in the bottom 50% were largely invisible to AI systems.
The mechanism is not complicated. AI engines decide what to cite based on which sources they've repeatedly indexed as credible — and the behavioral history of users who read those sources. Both inputs trace back to the same place: editorial trust built through earned media in publications that carry authority with the machines and with the humans they're personalizing for.
That's not a configuration problem. You can't fix it with structured data or FAQ schema. The correlation is to earned mentions — the kind that come from journalists and editors, not from content ops teams.
Why this matters more than the SEO piece
Personal Intelligence is now operating on consumer and B2B behavior simultaneously. Every founder, CMO, and VP of Growth who just got this feature turned on is walking around with an AI assistant that already knows which vendors they've heard good things about.
That assistant is going to surface those brands when they start researching solutions. According to Muck Rack's Generative Pulse analysis of over one million AI prompts, 85.5% of AI citations come from earned media sources. The AI doesn't invent preferences — it retrieves them from the editorial record it has indexed and that the user has encountered.
Which means the battle for brand authority wasn't lost yesterday when Personal Intelligence went free. It was lost three years ago, when a competitor was in Forbes and you weren't.
Forrester's research confirms the downstream consequence: 70% of B2B buyers complete most of their research before contacting sales. That research is increasingly AI-mediated. A buyer whose AI assistant already knows which vendors their professional network has read about and engaged with is not starting their evaluation from a blank slate. They're starting from a shortlist their AI built from their editorial history.
The TechCrunch coverage of yesterday's rollout noted that Google's AI can suggest a specific bag to match shoes a buyer just purchased — pulling brand preferences from past purchase history in Gmail. That's a consumer use case. The B2B version is: your prospect's AI assistant already knows which security software firms their colleagues have read coverage about in Wired or Dark Reading, and it will surface those firms when procurement starts.
If your company wasn't in those publications, no amount of on-site optimization changes what the AI recalls.
The common response will miss this entirely
When Google rolls out a feature like this, the marketing industry's first instinct is to ask: what can I optimize? How do I get into the recommendation pool through technical means?
The answers will come: structured data, entity optimization, schema markup, connected Google Business profiles. All of it is directionally correct and partially useful. None of it is the load-bearing variable.
The Princeton/Georgia Tech GEO research — the foundational academic work on AI citation behavior — found that adding statistics to content improves AI visibility by 30–40%. That's a content structure finding. But the prerequisite to any content structure optimization is distribution authority — being in the publications AI engines and human behavioral histories already index as credible.
The brands that Google's AI will recall as "preferred" for your buyers are the ones that built a consistent presence in Tier 1 publications over the years that shaped your buyers' reading habits. That record is fixed. Personal Intelligence just made it visible.
What changes, and what doesn't
Personal Intelligence accelerates a shift that was already happening. AI search was already personalizing results based on query context and behavioral signals. Yesterday's announcement took that personalization to every free-tier user and made the mechanism explicit: your AI knows which brands you trust because it knows what you've read.
For founders and growth executives trying to understand where brand authority actually comes from in this environment, the answer is the same as it was before: earned media in publications that buyers, analysts, and journalists treat as credible.
What changes is urgency. Personal Intelligence at free-tier scale means the behavioral trust signals that feed AI recommendations are now being actively compiled for a much larger pool of buyers. The brands that are consistently cited in trusted editorial sources — TechCrunch, Forbes, WSJ, industry-specific tier-one publications — are accumulating that signal at scale, across 75 million users' behavioral histories, right now.
The brands that are running content programs instead of earned media programs are accumulating nothing the AI can recall.
This is what Machine Relations defines as the new discipline: not just AI search optimization as a technical layer, but earned authority as the substrate that all personalization and retrieval runs on. PR's original mechanism — trusted editorial coverage in publications that shape professional opinion — is the same mechanism that now determines whether your brand shows up when a buyer's AI assistant constructs a recommendation. The publications haven't changed. The first reader has.
If your brand isn't in the publications that built your buyers' editorial intuitions, no personalization feature can surface you. Google can only recall what was already there.
Related Reading
- AI Visibility for Consumer Brands: The 2026 Earned Media Playbook
- AI Visibility for HR Tech Companies: How People Platforms Get Cited in Enterprise AI Search
See how your brand currently appears across AI search systems before the window narrows further: AuthorityTech Visibility Audit