Answer engine optimization: winning the citation, not just the ranking
Answer engine optimization (AEO) is the practice of making your business the source an AI answer cites. Where SEO earns a rank on a results page, AEO earns a mention inside the answer itself — in ChatGPT, Google AI Overviews and AI Mode, Perplexity, Gemini or Claude.
Where does AEO stand right now?
Search is shifting from earning a rank to earning a citation — and through mid-2026 it kept shifting, first to being the business the agent calls, now to being the product the agent buys. AI answers intercept a large share of queries before anyone clicks, but the brands quoted inside the answer come out ahead. Meanwhile the first peer-reviewed causal evidence has landed, and it is unkind to the tactic market: relevance and retrieval rank drive citations, and the clever text tricks stop working once everyone uses them.
Rank → citation → the agent's call. AI Overviews cut top-rank CTR ~58% (Ahrefs), yet brands cited inside the Overview get ~35% more organic clicks than rank #1 beneath it. Google's agent will now phone local businesses directly, and at I/O 2026 it added always-on information agents that scan the web without a new query.
Peer review deflates the tricks — and Google agrees. SIGIR 2026's 252,000-trial study finds topical relevance and list position dominate first-citation choice, and many popular GEO text edits are ineffective or harmful once multiple actors apply them. Google's own optimization guide mythbusts llms.txt, chunking and AI-specific schema: AI features run on core ranking systems.
The answer layer is fragmenting fast. ChatGPT's share of B2B AI referrals fell from 89% to 63% in eight months while Claude went 1.4% → 18.5% and Gemini quadrupled (Goodie). On raw usage, Similarweb has ChatGPT sliding ~76% → ~53%. Only 36 brands rank top-100 on every engine every month (Semrush). Pick your surfaces deliberately.
Commerce is moving inside the assistant. OpenAI opened a self-serve ChatGPT Ads Manager on July 22 with Best Buy and Lowe's live, while two rival agentic-checkout standards — ACP (OpenAI/Stripe) and UCP (Google/Shopify) — race to own the transaction. Feed hygiene and protocol support are becoming AEO problems, not just ecommerce ones.
Structure and freshness still win citations. Answer-first ~40-word blocks and question-first headings drive most citations — a definitive 1–2 sentence H2 opener is the strongest on-page factor (r=0.49). 83% of commercial-query citations go to pages updated in the last 12 months.
The entity layer is underpriced. Branded mentions remain the top correlate of AI visibility, and Wikidata presence explains 42–50% of variance in some verticals — while search appearances explain only ~5.8%. Most authority lives inside the model, not in your rankings.
What should a home or facility services company do about AI search?
local AEO
Local AI search is now the default, not the experiment. Reporting puts AI Overviews on roughly 48% of all searches and more than 80% of local service queries — up from about 20% in early 2024 — while consumer use of AI for local search jumped from ~6% in 2025 to ~45% in 2026. It is also no longer only about being named: Google's US agent will call local businesses on a homeowner's behalf, ChatGPT reads a licensed Yelp corpus of 330M reviews, and Business Profile management went live inside Gemini in July. Your listing, your review text, and how fast someone picks up the phone are all ranking surfaces now.
80%+
of local service queries now show an AI Overview, vs ~20% of all searches in early 2024
87%
of independent HVAC & plumbing contractors have effectively zero AI citation share in their own market — including firms with hundreds of 5-star reviews
45%
of consumers used AI for local search in 2026, up from ~6% in 2025
Your local-AEO priorities
Answer the phone, fast. Google's agent now calls contractors for homeowners — being reachable with a price band and an availability window is becoming a ranking-adjacent behaviour
Claim the citation gap. With ~87% of independent contractors at zero AI citation share, being the one local firm with structured, answerable content is still a land-grab
Complete & actively manage your Google Business Profile — categories, NAP, service area, description, posts, photos, reviews, Q&A
Connect your GBP to Gemini (live Jul 3, 2026, US, single-profile) to manage hours, review replies and posts and read customer query reports
Claim and polish your Yelp listing — reviews, ratings and photos now flow into ChatGPT answers, with Request-a-Quote inside the chat
List every service with a keyword-aligned description and a price range — these feed AI Overview answers directly
Build a unique service-area page for each city / ZIP you cover
Reply to reviews — including negative ones. AI reads the review text and synthesizes sentiment, and visible, accountable replies shape how it describes you
Publish cost guides, FAQs and seasonal maintenance content, refreshed at least annually — freshness is a citation factor
Add LocalBusiness + Service + FAQ schema that mirrors visible content
Where AI looks for you
Google Business Profile— the primary source for local AI answers; now editable from Gemini
Yelp— licensed into ChatGPT; reviews, photos & quotes
BuildZoom— license, permit & project verification
Angi & HomeAdvisor— homeowner review aggregators
Houzz— remodeling & design project proof
BBB— trust & accreditation
Keep your name, address & phone identical everywhere. Conflicting phone numbers across directories are a top reason businesses vanish from AI recommendations — and now, a top reason the agent dials the wrong number.
The checklist below is ordered by strength of 2026 evidence, not by how novel it sounds. Most of it is unglamorous.
AEO quick-start checklist
The moves with the strongest 2026 evidence.
Fix the underlying organic rank first — AI Overview citations come almost exclusively from the classic top-10, and retrieval position beats text tricks
Lead every page with a direct, ~40-word answer (44% of citations come from the intro; H2-opener is the top on-page factor)
Structure content as real user questions (H2s = prompts)
Write self-contained passages an LLM can lift whole
Add FAQPage / HowTo / Article / Author schema that mirrors visible text (FAQ schema ≈ +38% citation rate)
Refresh commercial pages annually — 83% of citations go to pages updated within 12 months
Build entity consistency — same names, facts, bios everywhere; claim a Wikidata item
Earn branded mentions across third-party sites (Reddit +2.7×, Wikipedia +1.8×)
Serve server-side rendered HTML — many AI bots don't run JS
Track citations per engine — only ~11% of cited domains overlap between ChatGPT and Perplexity
Add Claude to your tracking — it went from 1.4% to 18.5% of B2B AI referrals in eight months
Check your CDN's crawler defaults before Sept 15 — allow ai-input even if you block ai-train
Skip llms.txt as a ranking play — 97% are never requested
The answer engines
Where you're competing to be cited.
G
AI Overviews single-pass, few sources; ~48% of all searches, 80%+ of local service queries
AM
Google AI Mode 1B+ monthly users; ~200 countries, 98 languages; now runs background agents
C
ChatGPT Search ~15 sources/response; Wikipedia-leaning; Yelp-licensed local; ads live
P
Perplexity ~9–22 citations/response; Reddit-leaning; highest ROI per citation
Ge
Gemini 3.5 Flash in Search & agentic calling; ~3 sources/response
Cl
Claude fastest-growing surface; 18.5% of B2B AI referrals, up from 1.4%
What has actually changed in AEO recently?
Every item below is a summary of a named, linked source, with a note on what it changes in practice. Newest platform news first, then the research, the techniques and the tools.
Platform news
Google replaces the search box with agents — AI Mode passes 1 billion monthly users
At I/O 2026 Google said AI Mode had crossed one billion monthly users roughly a year after launch, with query volume more than doubling every quarter, and that it now reaches close to 200 countries and 98 languages at no subscription cost. The bigger structural change: "information agents" that run 24/7 in the background, scanning the web and reasoning across sources to push updates without the user issuing a new query. Google also shipped a redesigned, dynamically expanding search box that accepts images and video and feeds straight into AI Mode, with Gemini 3.5 Flash as the default model.
Why it mattersBackground agents mean your content can be evaluated when nobody is searching — a monitoring loop, not a query-response loop. Practically: publish machine-readable change signals (dated updates, structured pricing and availability, clean feeds), because the thing checking on your category may never load your homepage.
Google's agent will now phone your shop — home repairs on the launch list
Google rolled out agentic calling in the US: a searcher asks for a service, Google's AI dials local businesses on their behalf, asks the questions the user typed, and returns a written summary comparing who answered, their price band and their availability. Home repairs, beauty and pet care were on the launch list, with Gemini 3.5 Flash as the default model. Separately, Google Business Profile management went live inside Gemini on July 3, 2026 (US, single-profile) — hours, review replies, posts and customer query reports.
Why it mattersThe local funnel now ends in a phone call, not a click, and speed of answer becomes a competitive variable alongside your listing. For service businesses this is the most consequential change of 2026: make sure calls are answered fast by someone who can quote a price range and a window, and keep the phone number in your GBP exactly right — the agent dials what it reads.
ChatGPT Ads Manager opens — and two rival checkout standards fight for the transaction
OpenAI launched self-serve "Advertise in ChatGPT" with an Ads Manager on July 22, 2026, with Best Buy, Lowe's and VistaPrint among the first advertisers; targeting keys off conversational context rather than search keywords, and product-feed ads are part of the offering. Underneath sits a standards fight: the Agentic Commerce Protocol (ACP) from OpenAI and Stripe versus the Universal Commerce Protocol (UCP) launched by Google and Shopify in January 2026 with 20+ backers — and by June 2026 Shopify had dropped approval requirements for UCP-based agents entirely.
Why it mattersTwo things follow. First, the assistant surface now has a paid lane, so organic citation share and paid presence need to be measured together. Second, if agents will transact on your behalf, your product feed, price accuracy and protocol support become AEO assets — an agent can only buy what it can parse.
Yelp licenses 330M reviews to OpenAI — local answers go from crawled to contracted
OpenAI struck a non-exclusive content-licensing deal giving ChatGPT access to Yelp's ~330 million cumulative reviews and 8M+ business listings, including ratings, photos and business details. ChatGPT shows Yelp branding and back-links, and surfaces Yelp's "Request a Quote" flow so users can message local providers without leaving the chat.
Why it mattersFor local queries, ChatGPT is no longer inferring who is good from scattered crawls — it is reading a licensed review corpus. If you serve a geography, your Yelp listing quality and review text are now a direct input to AI recommendations, and the lead can be captured inside ChatGPT.
The AI browser shakeup: Atlas sunsets, ChatGPT Work arrives, Claude for Chrome hits 10M
OpenAI launched ChatGPT Work on July 9, 2026 — an agent built for hours-long, multi-step projects — and in the same announcement confirmed it was sunsetting the ChatGPT Atlas browser on August 9, 2026. Meanwhile Claude for Chrome went from roughly 40,000 installs in December 2025 to over 10 million by June 2026, and Perplexity's Comet remains in the mix, leaving a three-way agentic-browser landscape rather than the standalone-browser race vendors pitched a year ago.
Why it mattersThe extension model beat the standalone browser: agents are showing up inside the browser people already use. For measurement, that means more sessions that never carry a normal referrer — audit your GA4 attribution for AI traffic before you conclude an engine isn't sending you anyone.
Cloudflare turns Pay Per Crawl into Pay Per Use — and new crawler defaults land Sept 15
Cloudflare extended its Content Signals format in robots.txt with a use parameter letting site owners declare how bots may reuse content: Immediate (no storage or reuse), Reference (indexing, excerpts, links back) or Full. In parallel its Pay Per Crawl marketplace is evolving into Pay Per Use, paying publishers when their content creates value — e.g. when it appears in an AI answer — rather than only when a bot fetches it. From September 15, 2026 Cloudflare's defaults will block "mixed-use" crawlers from any page hosting ads, giving AI companies until then to separate search crawlers from AI training and agent crawlers. Reporting puts the affected footprint at roughly 8.5% of the top web, with publishers blocking AI about five times more often than the open web average. Google has said Content-Signal has no support among major crawlers and is not an enforceable block.
Why it mattersA live negotiation over whether being crawled and being cited can be priced separately. Practically: decide now whether you want ai-input access (you probably do — that's the citation pipe) even if you block ai-train, and check what defaults your CDN will apply for you in September.
Google publishes its first official AI-optimization guide — and calls AEO/GEO "still SEO"
Google added an official Search Central page, "Optimizing your website for generative AI features on Google Search," covering unique content, local/shopping/image/video guidance, and early AI-agent guidance. Its mythbusting section states that llms.txt, content chunking, AI-specific rewriting and special AI schema are not needed — generative AI features run on core Search ranking and quality systems.
Why it mattersThe closest thing to a rulebook from the platform itself, and it deflates a lot of the tactic market. Read it as: keep doing excellent SEO, add extractable structure and entity clarity, and stop paying for AI-specific file formats that Google says it ignores.
SIGIR 2026: 252,000 controlled trials on what actually wins the first citation
"What Gets Cited: Competitive GEO in AI Answer Engines" (Vishwakarma et al., presented at ACM SIGIR 2026) built a two-document RAG testbed injecting exactly two candidate sources into model context and measuring which one the first citation marker points to. Across six LLMs it ran 252,000 trials over 18 content factors — content match, completeness, trustworthiness, readability, competitive standing, freshness — with brand anonymization and counterbalanced ordering to separate content effects from position bias. Mixed-effects models found topical relevance and list position are the biggest drivers, and that many popular competitive-SEO document edits are ineffective or actively harmful once multiple actors apply the same trick.
Why it mattersThe first properly controlled, peer-reviewed causal evidence in this space, and it points away from the tactic market. Improving the underlying retrieval ranking of the page beats surface-level conversational rewriting — and any "GEO hack" that works only while you're the only one using it is not a strategy.
Goodie Wave 2: ChatGPT's share of B2B AI referrals fell from 89% to 63% — Claude took most of it
Goodie's 2026 AI Search Market Share Report (Wave 2) pulled GA4 referrer data from an anonymized brand panel and triangulated it against Similarweb traffic across the top five AI surfaces (25.77B visits, Jan–Apr 2026) and SensorTower iOS rankings. Over eight months ChatGPT's share of B2B AI referrals dropped from 89% to 63%; Claude went from 1.4% to 18.5% to become the #2 source; Gemini quadrupled and Perplexity more than doubled. Wave 1 had measured 2.8M AI referral sessions across 41 brand sites. Similarweb's separate usage data shows the same direction on a different metric — ChatGPT's share of gen-AI web traffic sliding from ~76% to ~53% while Gemini passed a quarter.
Why it mattersIf your AI-visibility dashboard only covers ChatGPT, you are now blind to more than a third of the referral pie — and to the single fastest-growing surface. Note also that usage share and referral share diverge sharply, so treat any single vendor number as directional and track both.
680M citations analyzed: ChatGPT and Perplexity share only 11% of their cited domains
Averi's analysis of 680 million AI citations in early 2026 found that just 11% of cited domains appeared in both ChatGPT and Perplexity results — the two engines draw on largely different source pools. The engines also cite at very different rates and volumes: BrightEdge measured ~8.79 citations per Perplexity response, and Superlines measured a 15.43% citation rate for Perplexity versus 2.78% for ChatGPT. Despite far smaller referral volume, Perplexity is reported to deliver the highest ROI per citation earned.
Why it mattersThis is the hardest number against a single "AI SERP" mental model. Winning on one engine tells you almost nothing about the other, so budget for engine-specific source-building — and don't dismiss Perplexity on traffic volume alone when its per-citation value is the highest measured.
Ahrefs: 97% of llms.txt files got zero requests — 137,210 domains analyzed
Ahrefs analyzed 137,210 domains with traffic in May 2026. Of the ~38,000 with a valid llms.txt, 97% received zero requests for it that month. Of the requests that did land, 96% came from bots and most were non-AI: SEO audit tools 21%, unidentified 14%, Googlebot and similar 13%. AI bots were 19% of requests, dominated by coding agents (10%). Separately, no major AI company reads llms.txt in production, and adoption has effectively flatlined in the 7–10%-of-domains range.
Why it mattersThis closes the llms.txt question for search. It is a developer-documentation and coding-agent convention, not a citation lever. Combined with Google's mythbusting, spending time here instead of on page structure is a straightforward misallocation.
The AI Overviews CTR reality-check: ~58–61% drop — unless you're the one cited
Ahrefs measured a 58% CTR reduction on the top-ranking page for AI Overview keywords (Feb 2026). Seer Interactive measured organic CTR falling 61% (1.76% → 0.61%) and paid CTR 68% across 3,119 informational queries; Pew found an 8% click rate with AI Overviews vs 15% without. But brands cited inside the AI Overview see ~35% more organic clicks and ~91% more paid clicks than brands sitting at position one below it. CTR on AIO-present queries also rebounded from 1.3% (Dec 2025) to 2.4% (Feb 2026).
Why it mattersThe headline "AI killed clicks" number and the AEO business case are the same data point read from two sides. Ranking beneath an AI Overview is the losing position; being quoted inside it is a net gain. That asymmetry is the entire argument for the work.
Semrush 2026 AI Visibility Index: 126M prompts, only 36 brands win everywhere
Semrush analyzed 126 million US AI prompts (Jan–Apr 2026) across ChatGPT, Gemini, AI Mode and AI Overviews, covering 1,200+ brands in 22 verticals. Only 36 brands ranked top-100 on every platform every month. ChatGPT cites ~15 sources per response; Gemini ~3. And 45% of marketing leaders say they can't accurately measure AI visibility.
Why it mattersCross-platform dominance is rare — pick your surfaces. The 15-vs-3 citation gap means your link-earning strategy should differ sharply by engine.
Wikidata beats backlinks: most LLM authority lives inside the model
The Digital Bloom's 2026 LLM ranking-factors analysis found search-engine appearances were the strongest global external signal but explained only ~5.8% of variance in LLM recommendation scores — 80–85% comes from inside the model. Wikidata presence dominated in several verticals (Hotels R² 42.3%, ERP software 42.9%, Furniture 49.9%), and backlink counts correlated via citation breadth and category fit rather than raw PageRank. Branded web mentions remain the highest-correlating metric for appearing in AI Overviews (Ahrefs).
Why it mattersThe highest-leverage, least-crowded AEO move right now is the entity layer: a clean Wikidata item, consistent structured facts, and broad third-party mentions. Almost no B2B brands have invested here, and rankings alone barely move the needle.
Answer-first blocks, question-first headings, and the 40-word rule
Lead each page or section with a clean, self-contained answer of roughly 40 words, then expand with evidence below. Turn headings into the exact questions people ask an AI ("How much does a frameless shower door cost?") and answer immediately underneath. Data: 44.2% of LLM citations come from the first 30% of a page's text, and a definitive 1–2 sentence H2 opener is the single strongest on-page citation factor (r=0.49).
Why it mattersThis is the single most repeatable AEO tactic in the 2026 studies — prompt-matched headings plus extractable, front-loaded answers get lifted verbatim into AI responses.
Freshness is a ranking factor for commercial queries
For commercial and evaluation-stage queries, 83% of AI citations came from pages updated within the past 12 months, and more than 60% from pages refreshed within the last six. Visible year signals in titles and headings ("… in 2026") are associated with roughly 30% higher citation rates.
Why it mattersYour best-performing pages decay silently. Put commercial and comparison content on an annual (ideally semi-annual) refresh cadence, update the visible date and year signal, and re-verify every stat — a stale page is invisible to engines even when it still ranks.
Schema that reflects visible content — FAQ, HowTo, Article, Author
FAQPage, HowTo, Article, Organization and Author/Person schema have the most impact when they mirror on-page text and reinforce authorship, entities and page intent. Keep FAQ answers conversational and under ~300 characters. Data point: complete, valid FAQ schema is associated with a ~38% higher citation rate, and sources with clear author attribution, publication dates and structured data have been measured as cited around 3× more often than those without. Note Google's caveat — no AI-specific schema is required for its generative features.
Why it mattersStructured data makes your entities and answers machine-legible — a core LLMO requirement. Use standard schema well rather than chasing AI-specific markup.
Entity mapping over keyword targeting + server-side rendering
Map the entities (people, products, places, concepts) your brand should own and keep them consistent everywhere — website, directories, Wikidata, third-party profiles. Optimize with semantic HTML and structured data, and serve server-side-rendered HTML because many LLM crawlers can't execute JavaScript.
Why it mattersIf a bot can't render your JS, your content simply doesn't exist to it. Entity consistency is how models decide you're the authority — and the 2026 data says that layer explains far more variance than rankings do.
GEO went from a fringe idea to a ~22,000-searches-a-month category in 2026, and the tooling has consolidated into tiers: enterprise visibility analytics (Profound, now with prompt volumes and shopping visibility), incumbent SEO suites bolting AI on (Ahrefs Brand Radar with a 243M+ prompt database across five engines including Claude and Meta AI; Semrush surfacing AI mentions beside organic rankings across four engines), lower-cost citation trackers (LLM Pulse, Otterly, SE Visible), full-stack GEO platforms (Goodie AI, from ~$495/mo), brand-level model analysis (Evertune) and content engineering (AthenaHQ, Writesonic).
Why it mattersPick by job-to-be-done: measurement first (are we cited, where, vs whom), then a content workflow to close the gaps. Check engine coverage before you buy — with Claude now a top-two referral source, a tool that only tracks ChatGPT, Gemini and Perplexity has a hole in it. The category is young; avoid long contracts.
Short answers, each one grounded in a source cited above.
Is AEO actually different from SEO?
Less than the tactic market wants you to think. Google's own generative-AI optimization guide says AI features run on its core ranking systems, AI Overview citations come almost exclusively from the classic top ten, and SIGIR 2026's 252,000-trial study found topical relevance and retrieval position dominate which source gets cited first. Fix the underlying organic rank before anything else — AEO is a layer on top of SEO, not a replacement for it.
Does an llms.txt file help you get cited?
Not as a ranking play. Ahrefs found 97% of llms.txt files were never requested at all across 137,000 domains, and Google's optimization guide explicitly mythbusts llms.txt alongside chunking and AI-specific schema. Publishing one is cheap and harmless, but it should not be counted as AEO work.
Do AI answers actually cost you traffic?
Both things are true at once. AI Overviews cut click-through on the top-ranking page by roughly 58% according to Ahrefs, but brands cited inside the Overview get about 35% more organic clicks than the page ranking #1 beneath it. The traffic loss is real; being named in the answer is what offsets it.
Which answer engines should a small business track?
More than one, and not the same ones forever. Only about 11% of cited domains overlap between ChatGPT and Perplexity, and just 36 brands rank in the top 100 on every engine every month. The mix moves quickly too: ChatGPT's share of B2B AI referrals fell from 89% to 63% in eight months while Claude went from 1.4% to 18.5% and Gemini quadrupled.
What single on-page change moves the needle most?
Lead with a direct answer of roughly 40 words underneath a question-shaped H2. Around 44% of citations come from a page's intro, a definitive one-to-two-sentence H2 opener is the strongest measured on-page factor (r=0.49), and FAQPage schema that mirrors the visible text correlates with roughly 38% higher citation rates. Serve it as server-rendered HTML, because many AI crawlers do not run JavaScript.
Why do home and facility service companies get left out of AI answers?
Because reviews alone do not do it. Roughly 87% of independent HVAC and plumbing contractors have effectively zero AI citation share in their own market, including firms with hundreds of five-star reviews. AI reads your Google Business Profile, a licensed Yelp corpus and the trade directories — and inconsistent name, address and phone details across those sources are a leading reason a business disappears from AI recommendations.
58%
CTR drop on the top-ranking page for queries with AI Overviews — Ahrefs, Feb 2026
+35%
more organic clicks for brands cited inside the AI Overview vs. rank #1 below it
1B+
monthly users on Google AI Mode one year after launch — Google I/O 2026
63%
ChatGPT's share of B2B AI referrals, down from 89% eight months earlier — Goodie, 2026
ChangelogAug 9, 2026 · refresh 5 — Added 5 new items (Google I/O 2026's autonomous information agents and AI Mode passing 1B monthly users; OpenAI's self-serve ChatGPT Ads Manager plus the ACP-vs-UCP agentic-commerce split; Goodie's Wave 2 finding that ChatGPT's share of B2B AI referrals fell 89%→63% while Claude hit 18.5%; Averi's 680M-citation overlap analysis; the AI-browser shakeup as Atlas sunsets and Claude for Chrome passes 10M installs); dropped 4 stale or superseded items; refreshed the stat band around AI Mode scale and referral share, and rebuilt the local-services stats around new AI-Overview prevalence data.
Earlier updates
Aug 3, 2026 · refresh 4Google's agentic calling began dialing contractors on a homeowner's behalf and Business Profile management went live inside Gemini; AI Mode gained the ability to act inside Instacart, Canva and YouTube; SIGIR 2026's peer-reviewed 252,000-trial citation study landed; Similarweb's engine-fragmentation data added.
Jul 28, 2026 · refresh 3Yelp licensed 330M reviews to OpenAI for ChatGPT local answers; AI Mode + AI Overviews launched in France plus 6 new languages; Google published its first official generative-AI optimization guide calling AEO/GEO "still SEO"; Ahrefs found 97% of llms.txt files got zero requests across 137K domains; Cloudflare shipped AI content-use controls; the AI Overviews CTR reality-check; and the Digital Bloom Wikidata ranking-factor study.
Jul 23, 2026 · refresh 2Google put Nano Banana image generation inside AI Overviews; Common Sense Media gave Google's AI Search its lowest safety rating for kids; Visionary's 8,400-prompt study (34% engine agreement on the top brand, FAQ schema ≈ +38% citation rate).
Jul 22, 2026 · refresh 1Added two studies: Semrush's 126M-prompt AI Visibility Index (only 36 brands rank top-100 across all four engines) and OtterlyAI (YouTube cited in ~16% of LLM answers, uncorrelated with views). Vault-backed from here on.
Jul 22, 2026Added a "For home & facility service companies" local-AEO section: Google Business Profile signals, key directories (BuildZoom, Angi, HomeAdvisor, Houzz, Yelp, BBB), NAP consistency, and per-city service-area pages.
Jul 22, 2026Initial feed published — 16 items plus a one-minute summary; weekly auto-refresh scheduled for Mondays.
Who put this together?
Carl Moore — thirty years in the cleaning and restoration trade, technician to franchise owner, now fixing the workflows that quietly slow service businesses down. This brief exists because the same owners keep asking the same question: why doesn't the AI ever mention us?
AEO moves fast. Every claim here links to its source — verify before you act on it. Figures are reported as published and are not independently audited. Tool listings are a directory, not endorsements.