AI Research Just Became Google's Default. The Brands Already in Those Answers Didn't Get There Through SEO.
Google expanded Canvas in AI Mode to all US users yesterday. When AI research becomes the default layer, the mechanism that puts brands in those answers rewrites the entire playbook.
Yesterday Google expanded Canvas in AI Mode to every English-speaking US user. Not a beta. Not a lab experiment. Every search user in the country now has access to an AI research assistant built directly into Google Search. Google said in its announcement that Canvas can draft documents, organize multi-source research, and turn raw notes into structured outputs — all inside the search interface.
TechCrunch's headline was technically correct: Google Search rolls out Gemini's Canvas in AI Mode to all US users. Accurate. Incomplete.
What actually happened yesterday: AI research became the default layer for the largest search platform on earth.
The shift has been building for two years. Google searches per US user fell nearly 20% year over year — the sharpest decline in the platform's history, driven by AI-powered answers absorbing queries before users ever reach a results page. ChatGPT now has 900 million weekly active users, a number OpenAI disclosed alongside a $110 billion raise. Google's Gemini app crossed 750 million monthly active users in February — a number that was 650 million last quarter. Gartner called this in early 2024: traditional search volume would drop 25% by 2026 as buyers shifted to AI. That prediction is landing on schedule.
Canvas accelerates the whole thing. A prospect who used to Google "best B2B PR firm" and scan ten blue links now describes what they need and watches AI organize an answer in real time. The research session isn't happening on your website. It's happening inside an AI interface that synthesizes what it knows about your category — and what it knows comes from what trusted publications have written about you.
That's the mechanism most brands are still missing.
Here's what AI search surfaces when someone researches your category: citations from publications it treats as authoritative. Forbes, Reuters, the Financial Times, TechCrunch, Axios. We mapped citation patterns across ChatGPT, Perplexity, and Google AI Overviews in March and found that Forbes is the only traditional media outlet cited by AI engines across all 11 major B2B and B2C sectors — consistently. Finance, tech, healthcare, SaaS, consumer. All of them.
The AI isn't checking your website first. It's consulting the editorial record that exists about you in the publications it trusts.
This is the part that doesn't fit the current marketing conversation. Brands are updating schema markup, building FAQ sections for AI Overviews, requesting llms.txt files. All of it matters at the margins. None of it is the primary lever.
The primary lever is whether your brand has been written about — credibly, recently, repeatedly — in the publications AI systems treat as trustworthy sources. Not once. Consistently. Earned media coverage in Forbes, the Financial Times, TechCrunch. That's what moves you from invisible to cited, and it compounds. A brand with 18 months of consistent placements in those outlets shows up in AI answers differently than one that landed a single feature two years ago.
Canvas makes this more consequential because it deepens the research session. A user running Canvas to evaluate vendors isn't skimming — they're building a structured view of the category. The brands that show up in that view get evaluated. The ones that don't, don't register.
The operational implication runs counter to where most marketing teams are focused right now.
SEO strategy — optimizing keywords, building domain authority through backlinks, structuring content for crawlers — was designed for a paradigm where Google returned a ranked list of URLs and users clicked through. That paradigm is being replaced at the top of the funnel. The buyer's first research pass is now AI-mediated, and what the AI says about your category is determined before the user types the query. It's downstream of the editorial record your brand built over time.
The brands already in Google's AI answers didn't get there through SEO. They got there because they have consistent earned media coverage in the publications AI engines index as authoritative. Some of it was intentional. Most wasn't — they ran traditional PR, landed in credible outlets, and are now benefiting from a model they didn't design for but happen to fit.
What's different now is that it's a choice. The mechanism is visible. The publications are known. The gap between brands that are in those answers and brands that aren't is going to compound fast as Canvas normalizes AI-assisted research across 300 million US search users.
This is what Machine Relations means as an operating discipline. It's the practice of ensuring that when AI systems research your category, your brand is in the sources they trust — through the same mechanism that has always determined editorial credibility: third-party placement in publications that earned AI systems' trust the same way they earned human readers' trust, over years of consistent, credible reporting.
PR's mechanism was always right. Earned media in trusted publications is the most durable trust signal that exists. The reader changed. The publications didn't. The brands that understand this are building presence in an AI search layer that Canvas just made permanent for every American with a Google account.
The brands that don't will spend 2026 optimizing for search behavior their buyers already moved on from.
If you want to see where your brand actually stands in AI answers right now, the audit takes three minutes. Most brands don't like what they find. Knowing the gap is how you start closing it.