AI Pulse: Google's First Official AI Search Guide Names What Actually Matters

Week of 2026-05-18 · Manta AEO

Google published its first authoritative guide to ranking in AI search this week, and the headline finding cuts against most of what agencies have been selling. The guide explicitly names three things you can ignore: llms.txt files, content chunking strategies, and elaborate schema markup. What it says to focus on instead is shorter, more practical, and easier to act on without a developer.

This Week in AI Search

Google’s Official AI Search Guide Says the Basics Are the Work

Google released its first official guide for appearing in AI Overviews, AI Mode, and Gemini results. The direct finding: llms.txt, special content chunking, and structured schema markup do not meaningfully affect how AI features surface your content. What does matter is factually accurate content, consistent business information, and a credible reputation — the same foundations that have always driven trust. For an independent chiropractor or optometrist in Ontario, this is clarifying: the fundamentals you can build without a development budget are the actual levers. Agencies selling technical AI-only add-ons have a harder argument to make this week. Google’s New AI Search Guide Calls AEO And GEO ‘Still SEO’

Google Analytics Now Tracks AI Assistant Traffic as Its Own Channel

Google rolled out a new default channel group in Google Analytics this week that separates traffic from AI assistants — ChatGPT, Gemini, Perplexity — from regular referral sources. Previously, AI citation visits landed in the “referral” bucket alongside everything else, making it impossible to see how much business AI platforms were actually sending. Any practice owner with Google Analytics can now see, for the first time, exactly how many people are arriving via AI citations. If that number is zero, that’s a clear signal. If it’s non-zero, the pages receiving those visits are your most AI-visible assets. Google Analytics Adds AI Assistant As Default Channel Group

AI Overviews Are Surfacing Negative Reviews Without Anyone Asking for Them

A new study documented that Google’s AI Overviews pull negative review content into search results even when the user’s query has nothing to do with reputation. A patient searching “dentist near Mississauga” may see a summary that surfaces a complaint from a one-star review — without typing anything about ratings. This is a structural shift in how reputation works. A single unresponded negative review used to affect a fraction of searchers who sorted by rating. AI Overviews can now move it to the top of a general query result. The fix is the same one it’s always been: respond to every review, including the negative ones. Data Shows AI Overviews Exposing Negative Reviews Without User Intent

Google began pulling social media content into Google Business Profiles this week via a new “Social Media Updates” carousel. When a business connects a social channel, recent posts appear on the profile page itself. For a physiotherapy clinic in Hamilton or a veterinary office in Kingston, this means Instagram and Facebook posts are now visible to anyone who finds the practice on Google. The AI visibility implication: Google’s data layer on your business just expanded. Practices that haven’t posted in months now have a visible gap on their own profile. Google is now pulling social media posts directly into Google Business Profiles

New Data on What’s Getting Cited

Ahrefs Research Finds Schema Has Less AI Citation Impact Than Expected

New research from Ahrefs, published alongside Google’s removal of FAQ schema from search results, found that structured schema markup has a weaker effect on AI citation rates than the optimization community assumed. This matters because a significant portion of AI search advice over the past year has centred on marking up FAQs, services, and business hours as the primary technical lever. The Ahrefs data, combined with Google’s own guide this week, makes the same argument: content authority and review volume drive citations more than schema additions do. For a solo accounting practice or independent immigration lawyer, that means less technical work is required — not more. SERP FAQ Removal & New Data Challenge Schema’s AI Search Value

AI Citation Traffic Converts at Higher Rates Than Most Channels

Ahrefs published analysis showing that visitors who arrive via an AI citation — a link inside a ChatGPT, Perplexity, or Claude response — convert at meaningfully higher rates than users from typical referral or social sources. The reason is intent: someone who asked an AI a specific question and clicked through a citation has already done more research than a casual browser. For an independent financial advisor or estate planning specialist in Ontario, one cited mention answering a specific client question is worth more than dozens of generic web visitors from a social post. The question is which pages are earning those citations. AI Chatbot Traffic: What It Is, and How to Get More

AI Visibility Is Three Separate Problems, Not One

Search Engine Journal published a detailed framework arguing that when a business is invisible to AI platforms, the failure almost always sits in one of three distinct layers: whether the AI knows the business exists, whether it trusts what it knows, and whether it retrieves the business for the right queries. Treating all three as one problem produces generic fixes that don’t move any of them. At Manta, audit results show this exact pattern: practices with reasonable citation counts but thin review profiles get passed over by AI platforms that have indexed them — because trust signals are missing, not awareness. Stop Treating AI Visibility As One Problem. It’s Actually Three, On Three Different Layers

What This Means for Independent Canadian Practices

Google’s confirmation that llms.txt and schema markup don’t drive AI rankings should redirect where practice owners spend time. The actual list from Google’s own guide is short: accurate content, consistent business information, and a credible review profile. A naturopath in Toronto or a solo family lawyer in Ottawa who has been told to wait for a developer to add schema should stop waiting. That technical work is not what moves the needle on AI citations right now.

The negative review exposure finding has immediate consequences for any regulated professional practice. Dentists, physiotherapists, and financial advisors operate in industries where reputation is the primary referral driver. AI Overviews amplifying unresponded complaints into general query results — not just review-specific searches — means review management is now front-of-practice work, not a background task. Responding to every review, including critical ones, is the most direct defense available.

The Google Business Profile social update is a fast action for any practice already active on Instagram or Facebook. Connecting a social channel takes minutes and automatically adds a content layer that AI systems read when building their picture of a business. A physiotherapy clinic that posts patient education content twice a week just had that content routed into their Google profile.

One Thing to Do This Week

Open Google Analytics and look for the new “AI Assistants” channel in your acquisition report. If you see visits — even a handful — check which pages they landed on. Those are the pages AI platforms already consider authoritative for your practice, and they tell you exactly where to invest more content effort. If the number is zero, that’s your baseline: AI platforms know you exist but aren’t sending anyone yet. Either finding takes five minutes to see and gives you the clearest picture of where you actually stand in AI search this week.

Sources