BizBuz

The AI Radar.

What ChatGPT, Perplexity, Claude and Google reward changes constantly. The AI Radar tracks the official sources and the best industry publications, and adapts our scan as soon as the rules shift — with proof attached. That way you know your report follows today's rules, not last year's.

89
signals picked up
89/89
quotes verified
32
sources tracked
4 October 2026
latest sweep
01

How the radar works

Every claim on this page travelled the same road:

  1. 01

    Track the sources

    We follow the official documentation of Google, OpenAI, Perplexity, Anthropic and Bing, plus the best industry sources such as Semrush, Ahrefs, Search Engine Land and Search Engine Journal.

  2. 02

    Literal quote required

    A signal only counts if it comes with a literal quote from the source — no summaries, no interpretations, no rumours.

  3. 03

    Independently verified

    Each quote is then re-fetched independently from the source and compared word for word. If it doesn't match exactly, it doesn't get in.

Only then do we decide: adapt the scan, note a confirmation, keep watching — or deliberately set it aside.

02

Checks adjusted

The world shifted, so our scan shifted with it. This is what we changed recently.

  • 15 June 2026Source: Google Search Central

    Google confirms: an llms.txt file has no effect at all on your visibility or ranking in Google.

    “these files aren't needed for Google Search (and won't negatively or positively impact your visibility or rankings)”

    Literal quote — Google Search Central

    What we did with it: We rewrote our llms.txt advice: it no longer produces an improvement point; we still mention it as an optional tip.

  • 8 May 2026Source: Google Search Central

    Since May 7, Google no longer shows FAQ blocks in search results. But Q&A on your site remains valuable: AI assistants pull their answers from it.

    “This feature will no longer appear in Google Search starting May 7, 2026.”

    Literal quote — Google Search Central

    What we did with it: We reframed our FAQ advice: no longer 'for the Google block', but as machine-readable Q&A that AI assistants quote from.

03

New check

Three AI platforms independently say the same thing: block their search crawlers and you don't exist in their answers. So the scan now checks this.

  • continuously validSource: OpenAI — crawlerdocumentatie

    If you block ChatGPT's search crawler, you disappear from ChatGPT's search answers.

    “Sites that are opted out of OAI-SearchBot will not be shown in ChatGPT search answers, though can still appear as navigational links.”

    Literal quote — OpenAI — crawlerdocumentatie

    What we did with it: New check added: we now verify that your site doesn't accidentally shut out the AI search crawlers of ChatGPT, Perplexity and Claude.

  • continuously validSource: Anthropic — crawlerdocumentatie

    Claude warns as well: blocking its search crawler makes your site less visible in Claude search results.

    “Disabling Claude-SearchBot on your site prevents our system from indexing your content for search optimization, which may reduce your site's visibility and accuracy in user search results.”

    Literal quote — Anthropic — crawlerdocumentatie

    What we did with it: Part of the same new check: three AI platforms independently say the same thing.

  • continuously validSource: Perplexity — crawlerdocumentatie

    Perplexity's search crawler needs access to show your site in Perplexity results — it is not used to train AI models.

    “PerplexityBot is designed to surface and link websites in search results on Perplexity. It is not used to crawl content for AI foundation models.”

    Literal quote — Perplexity — crawlerdocumentatie

    What we did with it: Part of the same new check.

  • continuously validSource: OpenAI — crawlerdocumentatie

    Search visibility and AI training are separate: you can allow the search crawler while refusing to let your content train AI models.

    “a webmaster can allow OAI-SearchBot in order to appear in search results while disallowing GPTBot to indicate that crawled content should not be used for training OpenAI's generative AI foundation models”

    Literal quote — OpenAI — crawlerdocumentatie

    What we did with it: That's why our check only penalizes blocking search crawlers. Blocking training crawlers is your right — we don't judge that.

  • 4 June 2026Source: Semrush

    Fix AI-crawler access first, then work on content: AI can't cite pages it can't reach.

    “A low score usually means your robots.txt or meta tags are blocking ChatGPT-User, OAI-SearchBot, or Google-Extended. Fix crawler access before content work. Because AI systems can't cite pages they can't reach.”

    Literal quote — Semrush

    What we did with it: Independent confirmation of the new check by a major SEO player.

04

Confirms our approach

Recent publications that back up what the scan already checks.

  • BrightLocal analysed over 200,000 local searches for 1,342 locations across 60 days. Google Maps mentions a tracked business in 66% of searches, AI answers in only 32 to 38% (ChatGPT 33%). Repeat the same question and ChatGPT returns on average only 23% of the same results. ChatGPT leans mainly on Bing and Yelp and hardly on the Google Business Profile. The study names no country; the examples are American, so the figures show a direction, not a Belgian fact.

    “While Google Business Profile is important for AI Mode and AIO, it's hardly used by ChatGPT.”

    Literal quote — BrightLocal — Local AI Visibility Study

    What we did with it: Confirms why we check your Bing Places listing as well as your Google profile. And a reminder that one AI answer is a snapshot, not a fixed ranking.

  • 15 September 2026Source: Cloudflare — AI-crawlerinstellingen

    Since September 15 (announced July 1) Cloudflare blocks AI training and agent bots by default for new domains, and only on pages that show ads; AI search bots stay allowed by default. Existing sites don't change on their own. The risk lies in your own setting: ticking 'block training' at Cloudflare can also lock out multi-purpose crawlers such as Googlebot and Bingbot. A small measurement (SeenSure, 1,046 sites) even saw search bots refused less often after September 15.

    “For all new domains onboarding to Cloudflare, the categories of Training and Agent will be blocked by default on the pages that display ads, while Search will remain allowed by default.”

    Literal quote — Cloudflare — AI-crawlerinstellingen

    What we did with it: Confirms our split between search bots (must get through) and training bots (your choice) when we check your bot settings.

  • 29 September 2026Source: HasData — AI Crawler Block Index

    HasData live-tested 2,096 websites with the real ChatGPT bot. 5.5% blocked GPTBot at their CDN or firewall while their robots.txt allowed it; conversely 39.5% of sites that ban GPTBot in robots.txt still let it in. Measured from a datacenter IP, top global sites, open dataset (CC BY 4.0). Lesson: robots.txt alone doesn't tell you whether an AI bot can actually read your site.

    “115 sites (5.5% of the tested subset) blocked GPTBot even though their robots.txt did not disallow it.”

    Literal quote — HasData — AI Crawler Block Index

    What we did with it: Confirms our robots.txt check and shows its limit. Candidate check (V1): actually request your site as an AI search bot to catch firewall or CDN blocks. Not in the scan yet.

  • 8 September 2026Source: DEV Community — Reese Calder

    A measurement of the 5,000 most visited websites (Tranco list, September 7) found that of the 535 sites blocking the training bot GPTBot, 238 (44.5%) also accidentally block OAI-SearchBot: the bot ChatGPT uses to cite sources. One author, one measurement, so indicative; a separate measurement on Cloudflare data confirms that overly broad robots.txt rules are a known risk.

    “535 sites block GPTBot. 238 of those, 44.5%, also block OAI-SearchBot.”

    Literal quote — DEV Community — Reese Calder

    What we did with it: Confirms our check on AI search bot access in robots.txt (P1-5-18): blocking 'all AI' often locks out the citation bot too.

  • Since August 2 Anthropic embeds, in newer Claude models, an invisible watermark in generated text and provenance data (C2PA) in files such as svg, png and jpg, worldwide, to comply with the EU AI Act (article 50(2)). Detection is for now only available to authorised bodies such as regulators, media and researchers; a customer or competitor can't check this themselves.

    “Marking will apply to output from supported models wherever Claude is offered, worldwide.”

    Literal quote — Anthropic — Claude Help Center

    What we did with it: No check: our scan judges texts on substance (facts, specificity), not on who wrote them. You may use AI for your website; make sure the facts are right.

  • 21 May 2026Source: Meta — crawlerdocumentatie

    Meta now says it explicitly too: allow their search crawler and Meta AI can cite you and link to you. Block it, and it can't.

    “Allowing Meta-WebIndexer in your robots.txt file helps us cite and link to your content in Meta AI's responses.”

    Literal quote — Meta — crawlerdocumentatie

    What we did with it: Meta added to our AI search crawler check — the fourth platform independently saying the same thing. Their training crawler is not held against you: blocking it is a legitimate choice.

  • 6 August 2026Source: Google Small Business Bulletin

    AI models look on your site for real, first-hand experience — not general text anyone could write.

    “Share unique experience over generic advice: AI models look for authentic, first-hand knowledge rather than general summaries that anyone can write.”

    Literal quote — Google Small Business Bulletin

    What we did with it: Confirms what we already measure: our checks on hollow words and on your own verifiable facts follow exactly this line.

  • 15 June 2026Source: Ahrefs

    Hard data on llms.txt: in a study of 137,000 sites, 97% of those files were never fetched by any AI bot.

    “97% of those files received zero traffic in May 2026. Nothing fetched them at all.”

    Literal quote — Ahrefs

    What we did with it: Confirms our earlier adjustment: llms.txt isn't on your work list. Put your energy into normal, readable pages.

  • continuously validSource: OpenAI Developer Platform

    OpenAI clarifies: block ChatGPT's search crawler and you vanish from the AI answers themselves — at best a bare link remains. Invisible where it matters.

    “Sites that are opted out of OAI-SearchBot will not be shown in ChatGPT search answers, though can still appear as navigational links.”

    Literal quote — OpenAI Developer Platform

    What we did with it: Confirms our crawler checks: we verify AI search bots are allowed in — especially with generic 'block all bots' rules that cause this by accident.

  • Google itself: you don't need special AI files or tricks to appear in Google and its AI answers.

    “You don't need to create new machine readable files, AI text files, markup, or Markdown to appear in Google Search (including its generative AI capabilities), as Google Search itself doesn't use them.”

    Literal quote — Google — AI-optimalisatiegids

    What we did with it: Confirms our course: no hype tactics in the scan, only what demonstrably works.

  • 16 June 2026Source: Semrush

    An FAQ with your customers' real questions is one of the biggest sources of AI answers.

    “Add an FAQ section that addresses the questions buyers actually ask, because FAQs are a major source of agent answers in this category and close wording matches how people phrase their queries”

    Literal quote — Semrush

    What we did with it: Confirms our FAQ checks and the reframed advice.

  • 20 January 2026Source: Ahrefs

    Many AI crawlers can't read business data if it only reaches the page via JavaScript — a hidden problem on many modern sites.

    “many AI crawlers can’t access schema data if it’s client-side rendered (e.g. via Javascript)—even though many sites rely on this setup.”

    Literal quote — Ahrefs

    What we did with it: Confirms our check that also looks at your site without JavaScript — the way AI crawlers do.

  • 28 August 2025Source: Ahrefs

    Fast, light, clearly structured sites are the easiest for AI engines to read and cite.

    “Sites that load quickly, avoid heavy JavaScript reliance, and maintain a logical internal structure make it easier for AI engines to parse information.”

    Literal quote — Ahrefs

    What we did with it: Confirms our checks on load time and page structure.

  • Clear headings and logical sections make your page navigable — for people and for AI.

    “People generally appreciate it when web pages are organized by paragraphs and sections, along with headings that provide a clear structure to navigate content.”

    Literal quote — Google — AI-optimalisatiegids

    What we did with it: Confirms our heading-structure checks.

05

On the radar

Developments we track before acting on them. Numbers that matter.

  • Since September 18 Google's new blocks for comparison sites (aggregators) and direct providers (suppliers) also show for local business searches, for users in the EEA, Belgium included. A sole trader or shop can appear as a supplier without a feed: Google takes the data from your website. The supplier block only appears when Google also shows the aggregator block; you can't force it.

    “The supplier unit is available to users in the EEA when the aggregator unit appears for queries related to hotels, flights, long-distance trains or buses, products, and local businesses.”

    Literal quote — Google Search Central — aggregator- en leveranciersblok

    What we did with it: No new check: the block draws on what Google can read from your website, which we already measure (indexing, structured data, contact and address details). We're watching whether it shows up in Belgium for local searches.

  • 24 September 2026Source: Google Search Status Dashboard

    On September 24 Google started the 'September 2026 spam update', worldwide and in all languages, so Dutch- and French-language sites too. The rollout can take up to two weeks (to around October 8); on October 4 it had not yet been reported as complete. It's a spam update, not a core update: only sites using spam techniques are affected.

    “Released the September 2026 spam update, which applies globally and to all languages. The rollout may take up to two weeks to complete.”

    Literal quote — Google Search Status Dashboard

    What we did with it: Nothing changed in the scan. If you see a sudden drop in those weeks, look first at mass-generated or copied pages.

  • OpenAI is expanding ChatGPT ads: the ChatGPT Ads app in the Shopify App Store, at first only for US merchants, comes from September 23 to every country where ChatGPT Ads are available (Belgium included, according to Belgian agencies). Ads Manager gets AI suggestions for ad copy and imagery based on your landing page. 'Sponsored Agents' are only being tested in the US.

    “The app will be available internationally in markets where ChatGPT Ads are available starting September 23.”

    Literal quote — OpenAI — Reimagining advertising with AI

    What we did with it: Nothing changed in the scan: paid visibility is separate from what we measure. Worth knowing: the AI suggestions start from your landing page, so a clear page helps here too.

  • Since September 24 Square connects to Apple Business (the successor to Apple Business Connect since April), so merchants keep their details on Apple Maps, Siri AI and Apple Wallet up to date from their till. The integration covers eight countries, not Belgium. Claiming an Apple Business listing yourself is possible in Belgium. Siri AI is not yet available in the EU on iPhone, iPad and Apple Watch. The '30% more views' for claimed listings is Apple's own figure.

    “Square sellers can now easily connect to Apple Business to manage their presence across Apple Maps, Siri AI, Apple Wallet, and other Apple apps, keeping their business details in sync from Square Dashboard”

    Literal quote — Square — Apple Business-integratie (Business Wire)

    What we did with it: We already check whether your website links to your Apple Maps listing. Candidate check (V2): verify whether your Apple Business listing is claimed and complete. Not in the scan yet.

  • 3 September 2026Source: OpenAI — GPT-6 Astra

    On September 3 OpenAI launched GPT-6 Astra, the new flagship model behind ChatGPT search, browsing and agent tasks. It's available to paying subscribers (Plus, Pro, Business, Enterprise) worldwide, Belgium included, not to free users. Astra finishes computer tasks almost twice as fast; the paid 'Fast mode' add-on isn't available with EU data residency.

    “Astra is state-of-the-art on computer use, browsing, software engineering, cybersecurity, science, and professional work.”

    Literal quote — OpenAI — GPT-6 Astra

    What we did with it: Nothing changed in the scan. We're watching whether the new model cites small businesses more often; our AI questions measure that on every scan.

  • For shopping questions ChatGPT increasingly recommends products through structured product feeds (Shopify among them) rather than by crawling your website: from 8% to over 60% of recommendations, measured on July 10. The study is US-based and drawn from one tool's customers, so the exact figures don't automatically apply to Belgium.

    “feed-integrated retrieval overtakes web-search on July 10th, rising from 8.26% to 61.54%.”

    Literal quote — Profound — ChatGPT Shopping-onderzoek

    What we did with it: Candidate check (V1): if you sell products, we'll check whether you have an active product feed and not just a website. Not in the scan yet.

  • In a new documentation page Google describes a results layout where a small supplier (say a holiday home or hotel) can appear directly next to a booking platform as a 'supplier unit'. For now this covers only hotels, flights, transport and product sales, not local shops or services in general, and requires structured data on your site. Available in the EEA, so in Belgium too.

    “Google is evolving the search results page for certain types of queries to provide new opportunities for Vertical Search Services (VSS), Comparison Shopping Services (CSS), direct suppliers, and content providers.”

    Literal quote — Google Search Central — aggregator features

    What we did with it: Our schema.org checks (pillar 1) already cover the basics. We're watching whether Google extends the list to other sectors.

  • 14 September 2026Source: Apple Newsroom

    At the iOS 27 launch on September 14 the new Siri AI is not available in the EU. Apple says EU rules (Digital Markets Act) would give any AI system near-unlimited access to your device and refuses on privacy and security grounds. No EU date; on Mac, Siri AI does work.

    “we do not currently have a timeline for Siri AI's availability on iOS and iPadOS in the EU.”

    Literal quote — Apple Newsroom

    What we did with it: No check needed: Belgian customers can't use Siri AI on their iPhone to find a business yet. We'll pick this up as soon as Apple gives an EU date.

  • Visa, Mastercard and Ant International are starting a 'Know Your Agent' collaboration: a way to recognise AI agents paying by card on a customer's behalf. It's not a finished standard yet, but an early collaboration without technical details or timeline. Nothing changes for your till or web shop today.

    “AI agents are evolving from making recommendations to completing purchases on behalf of users.”

    Literal quote — Ant International, Mastercard en Visa — persbericht

    What we did with it: Nothing changed. We're tracking this because it will determine whether an AI agent can soon check out with you without you doing anything.

  • Since September Microsoft Copilot shows visual answer cards, including for place questions (like 'hairdresser in Ghent'), for work and personal accounts, rolled out worldwide. Microsoft hasn't confirmed which data source (Bing Places, for instance) feeds those place cards.

    “Surface new model-driven, dynamically placed Answer Cards across top 7 segments: weather, sports, finance, images, video, places and news.”

    Literal quote — Microsoft 365 Message Center (RM567008)

    What we did with it: Our Bing Places checks (pillar 2) stay in place; it's the most likely source for such a card. We'll verify as soon as Microsoft names the data source.

  • 18 August 2026Source: OpenAI — ChatGPT Ads

    Since August 24, ChatGPT shows ads to free users in 31 European countries — Belgium included, though OpenAI doesn't name our country; Belgian agencies confirm the rollout. Since August 31 the self-service tool is open across Europe too, Belgium included; on September 4 audience and conversion tools were added (updated September 14, run 7).

    “Ads in ChatGPT are always clearly labeled and separate from ChatGPT's answers, and advertising does not influence the answers ChatGPT provides.”

    Literal quote — OpenAI — ChatGPT Ads

    What we did with it: Nothing changed in the scan: per OpenAI, advertising doesn't influence the answers. The self-service tool is now open, so this has become a channel for small businesses too; we're watching that AI answers stay separate from ads.

  • continuously validSource: OpenAI — crawlerdocumentatie

    Blocking ChatGPT's crawler in robots.txt is no guarantee: when a user asks about your site live, ChatGPT-User may visit anyway, per OpenAI. A measurement across European sites found this actually happens on nearly half the sites that block it.

    “Because these actions are initiated by a user, robots.txt rules may not apply.”

    Literal quote — OpenAI — crawlerdocumentatie

    What we did with it: Our advice stands: just allow the AI search crawlers. This signal mainly shows that blocking doesn't even do what people think it does.

  • 26 August 2026Source: Anthropic — Claude in Chrome

    AI now sits inside the browser too: on any paid plan, Claude can click, type and fill out forms autonomously — no per-step approval. No European carve-out: this works in Belgium as well.

    “It can view the page you're on and take actions like reading and typing text, clicking links, navigating between pages, and filling out forms, using your existing logins.”

    Literal quote — Anthropic — Claude in Chrome

    What we did with it: Candidate check registered: your contact or booking form must be fillable by such a browser agent too — no aggressive bot-blocking or puzzles on the form itself.

  • 20 August 2026Source: Google Search Central

    Google now offers a button you add with two lines of HTML: visitors mark your business as a 'preferred source' and see you more often, including in AI answers. Condition: your site publishes fresh content regularly — a static site gains nothing. Works in all languages, here too.

    “By adding only two lines to your HTML, you can add an automatically translated, interactive button, which lets readers seamlessly add your site as a preferred source and return to where they left off on your page.”

    Literal quote — Google Search Central

    What we did with it: Candidate check registered for businesses with a news or blog page. For sites without fresh content this is not an improvement point.

  • 28 August 2026Source: Google Search Central

    Google for the first time lists which file formats work for your site icon (favicon) in search results — SVG isn't on it. The docs previously pointed vaguely to 'any valid format'. Google separately confirmed nothing changed functionally: the documentation is simply honest now.

    “Google Search supports the following favicon file formats: BMP, GIF, ICO, PNG, JPEG, PPM, and TIFF.”

    Literal quote — Google Search Central

    What we did with it: Our advice: use PNG or ICO for your site icon, at least 8x8 and preferably larger than 48x48 pixels. With an SVG icon, Google shows a grey dot next to your name.

  • 28 August 2026Source: Google Search Central Blog

    Google is softening — following discussion with the European Commission, Google itself says — its penalty for 'borrowed reputation' (paid guest content on someone else's strong domain): inside the EEA the manual action no longer applies as of August 30, though the affected site section can still be ranked separately.

    “For users inside the EEA, the impact of the manual action won't apply.”

    Literal quote — Google Search Central Blog

    What we did with it: No impact on our checks — this mainly affects large publishers renting out their domain, not the self-employed owner of their own site. We track it as a sign that European search results are getting rules of their own.

  • 25 August 2026Source: Carrefour België — persbericht

    Carrefour Belgium is the first Belgian retailer you can order from directly through ChatGPT: build a shopping list and week menu, check availability and order, via partner Mealz.ai. AI commerce is no longer an American story — it now runs in Belgium.

    “les clients peuvent désormais faire leurs courses directement depuis ChatGPT”

    Literal quote — Carrefour België — persbericht

    What we did with it: Nothing to do yourself today: this runs through a tech partner for large players. We're tracking when such integrations open up to small businesses.

  • continuously validSource: Facebook — robots.txt

    Facebook doesn't shut AI crawlers out — they're allowed in. But they find an empty box: from a public Facebook page, an AI crawler receives only the name and the town.

    “User-agent: GPTBot / Disallow: /*/plugins/* / Disallow: /?*next= / Disallow: /a/bz?”

    Literal quote — Facebook — robots.txt

    What we did with it: Measured by us on 13/08: a public Facebook page yields three visible words to an AI crawler — everything else sits behind JavaScript. Our explanation on the 'no website' page was right in its conclusion but wrong in its reason, and has been rewritten.

  • 24 July 2026Source: Google Search Central

    Since 24 July, Google explicitly bans reviews you gave customers something for without disclosing it. Do it anyway and Google can remove your stars from search results.

    “Don't include fake or undisclosed incentivized reviews on your page or in your structured data markup.”

    Literal quote — Google Search Central

    What we did with it: On our list as a candidate check: if your site shows reviews with stars, the disclosure belongs there. Not built yet — we don't add work to your list hastily.

  • 6 August 2026Source: Google Small Business Bulletin

    Google now says it in plain words: put your hours, services and menu as normal text on your site. Inside a photo, an AI cannot read them.

    “Type your hours, services, and menus directly as readable text on your website. Don't bury them inside flat photos or image graphics that search engines can't read.”

    Literal quote — Google Small Business Bulletin

    What we did with it: Top of our list as a candidate check: a menu or opening hours that exist only as an image are invisible to AI. We'll build it only after a design that avoids false alarms.

  • 10 August 2026Source: OpenAI — ChatGPT Release Notes

    Since 10 August, ChatGPT shows free tables and lets people book right away. Globally that runs through OpenTable only — if you're not on it, you don't exist in that conversation.

    “Reservations are available globally with OpenTable and in the US with Resy and in the US and Canada with Yelp.”

    Literal quote — OpenAI — ChatGPT Release Notes

    What we did with it: Noted as a channel outside your own website. In Belgium few businesses are on OpenTable, so this doesn't produce an improvement point; we mention it as a tip for hospitality.

  • Since 2 August, a chatbot must make clear it isn't human. That duty falls on whoever builds or sells the chatbot, not on you if you use an off-the-shelf tool.

    “Providers of AI systems that directly interact with people must design and develop those systems in such a way that the individuals concerned are informed that they are interacting with an AI system, unless this is obvious.”

    Literal quote — Europese Commissie — AI Act-FAQ

    What we did with it: Purely informational, not a check and not legal advice: do check whether the chat widget on your site says itself that it's AI. The official explanation is with the European Commission.

  • 29 July 2026Source: Similarweb

    Even after a year of growth, only about 7 in 100 US ChatGPT answers contain a cited source.

    “Citation presence in US ChatGPT prompts rose from about 1.6% in June 2025 to roughly 6.8% by May 2026.”

    Literal quote — Similarweb

    What we did with it: Important context for your result: zero mentions doesn't mean you're invisible — cited sources are rare anyway, and this figure covers the US.

  • 6 August 2026Source: Yelp

    Yelp says its reviews now power local recommendations in ChatGPT. OpenAI confirms this nowhere itself.

    “our trusted content is powering local discovery for ChatGPT and other AI partners”

    Literal quote — Yelp

    What we did with it: On the radar: if AI assistants license reviews from platforms, your profile there weighs more than your own text. Yelp is marginal in Belgium, so no action for now.

  • ChatGPT now deliberately uses your location to surface nearby businesses, combining business info and images into a single answer.

    “Shopping and local business queries should also feel more useful. The model makes better use of location context to surface relevant nearby options and can bring together product recommendations, business information, and images more coherently when those elements are helpful.”

    Literal quote — OpenAI — ChatGPT Release Notes

    What we did with it: Underlines why your business data must be correct everywhere: ChatGPT builds its local recommendations on what it finds about your business.

  • 8 July 2026Source: Search Engine Land

    Research: ChatGPT switches between hidden search pipelines behind the scenes — which shapes exactly which sites get cited. Visible today doesn't mean visible tomorrow.

    “Your page may never be considered if ChatGPT skips search, uses another retrieval source, or finds a clearer third-party page to support the claim.”

    Literal quote — Search Engine Land

    What we did with it: Strengthens our advice to be broadly present: your own site, your Google profile and review sites — never lean on one channel.

  • Google now calls small businesses itself with an AI assistant to verify practical details (parking, pickup, walk-ins) — and shows that info in Search and Maps. A new 'Collected info' screen lets you review it.

    “To help local customers find you, Google may occasionally call your business using an automated assistant to verify everyday details—like whether you offer curbside pickup, have dedicated customer parking, or accept walk-ins.”

    Literal quote — Google Small Business Bulletin

    What we did with it: On the radar as a candidate check: is the 'collected info' on your Google profile correct? Take those AI calls seriously — your answers land straight in Google.

  • 7 July 2026Source: Google Search Central Blog

    Search Console now also measures how your social and video posts perform in Google — even if your business has no website of its own ('platform properties').

    “we're thrilled to introduce platform properties, a new Search Console property type to help site owners and creators understand how their social and video posts perform on Google Search and Discover.”

    Literal quote — Google Search Central Blog

    What we did with it: On the radar: relevant for businesses that only live on Instagram or Facebook. Your own site remains our advice — this is a measuring tool, not a replacement.

  • 9 July 2026Source: TechCrunch

    OpenAI is shutting down its separate AI browser and moving that power into a ChatGPT extension for Chrome: AI that reads along live and performs tasks on the pages you visit.

    “OpenAI is launching a ChatGPT extension on Chrome that gives it access to the context of the page you're viewing, lets users ask questions about web pages, summarize content, or start longer tasks all from the browser.”

    Literal quote — TechCrunch

    What we did with it: On the radar: AI agents will soon read your site in your customer's ordinary browser. Clear, factual pages aren't just for crawlers — also for AI reading along live.

  • 30 June 2026Source: Semrush

    Same question, different reasoning mode: between ChatGPT's fast and deep modes only 25% of cited sources overlap. Three in four citations differ.

    “For the same prompt, only 25.6% of cited domains are shared between minimal- and high-reasoning modes. Almost three in four cited sources are different.”

    Literal quote — Semrush

    What we did with it: Reinforces the same lesson as the pipeline study: AI visibility is fluid. Broad presence and regular rescans are the only stable strategy.

  • 2 July 2026Source: Which?

    Consumer watchdog Which?: Tripadvisor's AI review summaries can bury serious complaints. AI summarises your reputation — and that summary is what people read.

    “Tripadvisor's new AI review summaries are masking reports of food poisoning, sexual harassment and serious hygiene failures, a Which? investigation has found.”

    Literal quote — Which?

    What we did with it: On the radar: review platforms let AI summarise your reputation. Fresh, genuine, plentiful reviews weigh double — they feed the summary.

  • 16 June 2026Source: Bing Search Blog

    Bing Webmaster Tools now shows your 'Citation Share': the share of AI citations your site captures — with 'local' as its own category. AI visibility becomes measurable.

    “Citation Share shows how much of the citation space your site receives for a specific grounding query. It is calculated as the percentage of citations attributed to your site out of all citations Copilot generated for that query.”

    Literal quote — Bing Search Blog

    What we did with it: Confirms how much Bing matters for ChatGPT/Copilot visibility, and sits on the radar as a candidate check: tracking your Citation Share in Bing Webmaster Tools.

  • 30 June 2026Source: Search Engine Journal

    For local questions, ChatGPT shows only two businesses. You're in the top 2, or you don't exist.

    “Ask for the best coffee near you, and ChatGPT returns two places, not a top 10. For local, you’re in the top 2, or you aren’t there.”

    Literal quote — Search Engine Journal

    What we did with it: On the radar: we track how AI assistants select local businesses.

  • 19 May 2026Source: Search Engine Land

    63% of nearly 400 million AI citations point to listicles ('best X in Y'). Being in those lists is gold.

    “Of the roughly 25,000 unique URLs we reviewed, half were listicles. Across nearly 400 million citations from all models, 63% pointed to listicles.”

    Literal quote — Search Engine Land

    What we did with it: On the radar: possibly future advice on presence in local list articles.

  • 10 June 2026Source: Search Engine Land

    16% of all AI prompts are location-based: 'near me' searching has migrated from Google to AI assistants.

    “16% of prompts are explicitly location-based. The “near me” query pattern has successfully migrated from Google to LLMs.”

    Literal quote — Search Engine Land

    What we did with it: On the radar: underlines why local AI findability matters now.

  • 6 April 2026Source: Search Engine Land

    Not Google but Bing largely determines which brands ChatGPT mentions. Ignore Bing and you miss ChatGPT.

    “You can dominate query fanouts on Google SERPs and still underperform in ChatGPT brand mentions.”

    Literal quote — Search Engine Land

    What we did with it: Confirms why our scan also checks your Bing presence.

06

Deliberately not adopted

Not every hype becomes a check. What we set aside — and why — belongs here too.

  • Brand mentions across the web are said to be THE big AI factor — but Google warns that inauthentic mentions don't help.

    “seeking inauthentic "mentions" across the web isn't as helpful as it might seem.”

    Literal quote — Google — AI-optimalisatiegids

    What we did with it: Deliberately no 'count your mentions' check: mention authenticity can't be measured from a site scan, and inflating them artificially doesn't help.

  • Google adds nuance: structured data isn't required for AI search — but remains useful for regular results and helps machines recognize facts.

    “Structured data isn't required for generative AI search, and there's no special schema.org markup you need to add. However, it's a good idea to continue using it as part of your overall SEO strategy, as it helps with being eligible for rich results on Google Search.”

    Literal quote — Google — AI-optimalisatiegids

    What we did with it: Noted, nothing changed hastily: other sources show structured data does help AI recognize business facts. We'll review the weighting deliberately, not reactively.

  • 15 May 2026Source: Google Search Central

    Google's spam policies also apply to AI answers in search results.

    “the spam policies apply to all of Google Search, including generative AI responses”

    Literal quote — Google Search Central

    What we did with it: No separate check needed: this simply confirms that tricks don't pay off in the AI era either.

Is your site ready for today's rules?

The scan checks your site against the current state of this radar — in a few minutes.