Siri AI Arrives September 14: How B2B Brands Get Cited in Apple's New Answers

Siri AI arrived with iOS 27 on September 14, 2026, as an English beta with usage limits. Apple's crawler documentation says pages collected by Applebot can help answer broad questions in Siri and Search, and those answers may link to their sources. For a B2B brand, eligibility starts with Applebot reaching your pages and no nosnippet on the ones you want used. Apple offers no citation reporting.

Below: what launched and where, what Apple confirms about sources, a six-step checklist, and how to measure a surface that sends you no data. For the wider crawler landscape, start with how AI crawlers work.

What launched on September 14, and where

At its September 9 event, Apple said iOS 27 would ship on Monday, September 14, alongside iPadOS 27, watchOS 27 and macOS 27, Engadget reports. Siri AI is the headline feature. When Apple introduced Siri AI in June, it said the assistant would be available as a beta for supported devices set to English.

• Languages. English on September 14. Apple's Lilian Rincon said French, Japanese, Korean, Portuguese and Spanish follow in October, as 9to5Mac reported.
• Devices. iPhone 15 Pro and later, the latest iPad mini, and iPad Air and iPad Pro models with M1 or newer, per Engadget. On iPad and Mac, Apple says Siri AI is integrated into Spotlight.
• Markets. Apple says Siri AI is not available initially in the EU on iOS, iPadOS and watchOS, though EU users on Mac and Apple Vision Pro can use it in a supported language. It is not available in China while Apple works through regulatory requirements.
• Usage limits. Apple lists Siri AI among the features that rely on server-side models and have usage limits, which MacRumors reports are daily. Apple says limits vary by feature, request complexity, system demand and other factors, that a feature becomes available again after a time out period, and that increased access will be available for a fee.
• Models. Apple describes its next generation of Apple Foundation Models as custom-built in collaboration with Google and its Gemini models. Engadget reports that Apple used Gemini models in training the algorithms behind Siri AI.
• Scale. At the event, Rincon said Siri users make over 2.5 billion requests every day, Search Engine Roundtable noted. That figure covers all Siri use, not web answers alone.

Why executive buyers on iPhone matter for B2B

For B2B teams selling into the US, Canada, India, Singapore or the GCC, none of those markets is on Apple's stated exclusion list, and English is the launch language.

We will not quote an iPhone share for senior executives, because we have not found a figure we trust, and the argument does not need one. Picture a CIO between board sessions asking Siri which firms implement a platform in the Gulf. That question never reaches your Search Console, and if the buyer does not tap a link, there is no visit in GA4. If Siri names two competitors and not you, nobody on your team will see it.

How Siri AI finds and cites sources

Apple's About Applebot page, published September 4, 2026, is the primary document on how web content reaches Apple's AI answers. It establishes four points.

1. Applebot is the crawler that supplies web content. Its data powers search in Spotlight, Siri and Safari, and may provide additional context and up-to-date content when AI models generate output in Apple products. Apple's example is answering broad world knowledge questions in Siri and Search, answers that "may include links to sources and websites used to help generate the answer."
2. nosnippet is the opt-out for answers. Applebot will not generate a description or web answer for a page tagged nosnippet, and Apple will not use that content as context for AI output. Apple supports it as a robots meta tag, a meta tag named applebot, or an X-Robots-Tag header.
3. Applebot-Extended is a training control. It does not crawl. Disallowing it opts your content out of training Apple's foundation models, pages that disallow it can still appear in search results, and Apple says its rules are not considered in ranking for Search.
4. Paywall markup also removes AI use. Pages marked isAccessibleForFree: false stay eligible for search results, but Apple will not use them as context for AI output. B2B teams that mark gated reports this way should know the trade.

Two more details catch B2B sites. If robots.txt does not mention Applebot but does mention Googlebot, Applebot follows the Googlebot rules. And Applebot may render pages in a browser, so JavaScript or CSS blocked in robots.txt can leave it with an incomplete page.

What Apple does not document

The Gemini connection invites a shortcut: if Gemini cites us, Siri will too. Apple has not documented that. The collaboration Apple describes concerns its foundation models, and the Applebot page names Applebot as the source of web context. Treat Siri AI as its own surface with its own crawler, and do not report Gemini results as a stand-in for Siri. Our piece on cross-platform AI visibility explains why results already diverge between engines.

What B2B teams control: a six-step checklist

1. Allow Applebot in robots.txt, and in CDN and WAF bot rules.
2. Scan templates, meta tags and headers for stray nosnippet.
3. Decide on Applebot-Extended as a separate training question.
4. Structure key pages with direct answer blocks and entity facts.
5. Align company facts across your site, profiles and directories.
6. Check Cloudflare's crawler settings before September 15.

Step 1: Allow Applebot at every layer

robots.txt is the first layer, not the only one. Bot management rules at the CDN or web application firewall can block a crawler that robots.txt allows. Check server or CDN logs for Applebot requests, and confirm them the way Apple describes: reverse DNS in the applebot.apple.com domain, or a match against Apple's published IP ranges. A request that carries the Applebot user agent from any other address is not Applebot.

# robots.txt: name Applebot explicitly
User-agent: Applebot
Allow: /

# Confirm a logged IP resolves to applebot.apple.com
host 17.58.101.179

If a firewall rule or challenge page sits in front of your marketing pages, exempt verified Applebot traffic as you would Googlebot.

Step 2: Remove stray nosnippet

nosnippet usually arrives by accident: a CMS option applied to a whole template, an SEO plugin setting, or a header added at the server for PDFs. Search the page source and response headers of your service, industry and comparison pages for nosnippet, including <meta name="applebot"> and X-Robots-Tag: applebot: nosnippet. Keep it only where you have decided a page should not feed AI answers. While you are there, confirm isAccessibleForFree: false sits only on content that is truly gated.

Step 3: Decide on Applebot-Extended separately

Apple describes Applebot-Extended as the way to opt out of training its foundation models, and names nosnippet, not Applebot-Extended, as the way out of AI answers. That makes training a legal and brand policy question, separate from answer eligibility. Make the call with legal, write it down with an owner, and do not let a training block get bundled into a broader "block AI" rule that also catches Applebot.

Step 4: Make pages citable

Apple publishes no guidance on how it picks sources for AI answers, so this step rests on what makes a page quotable for any answer engine, not on Apple documentation. Open each key page with a 50 to 70 word answer to the question the page targets. State entity facts in plain sentences: what the company does, for whom, where it operates, which platforms it works with, and named proof. Keep that text in the raw HTML, not loaded by scripts after the page renders.

Specifics matter by sector. A system integrator should say which platforms it implements and in which regions. A healthtech vendor should name the standards and systems it integrates with. When we rebuilt Ventive's search presence around the specific services and industries its buyers look for, the engagement added $800,000 of revenue in nine months. This is the core of our AEO, GEO and SEO work.

Step 5: Keep company facts consistent

Conflicting descriptions give any answer engine a reason to hedge or to pick the wrong version. Align your company name, one-sentence description, category, headquarters and markets across your site, LinkedIn, review profiles, partner directories and press boilerplate. Do the same for leaders: executive bios on your site, LinkedIn and conference pages should agree on title, company and expertise, which is part of CXO branding. Regional offices, such as a Singapore branch, should match the global site in local listings, which is part of our Singapore AEO and GEO work.

Step 6: Check Cloudflare before September 15

Cloudflare now sorts AI crawlers into Search, Agent and Training. From September 15, new customers and new sites get defaults that block Training and Agent crawlers on pages that display ads, while Search stays allowed, and existing free customers who have not changed their settings move to those defaults, Search Engine Journal reports. Cloudflare's announcement names Applebot, with Googlebot and Bingbot, as multi-purpose crawlers that will be blocked for customers who choose to block Training. A B2B site that blocked Training to keep its content out of models can lose Applebot too. Our walkthrough of Cloudflare AI crawler settings covers the dashboard.

Measuring what you cannot see

Apple's documentation describes no citation report, no impression data and no referral parameter. If Siri names your company in a spoken answer and the buyer never taps a link, nothing reaches your analytics. Every method below is indirect. Used together and against a baseline, they can show that something is changing. They cannot prove Siri caused it.

1. Branded search trends. Record branded query impressions in Search Console and branded interest in Google Trends for the weeks before September 14. A buyer who hears your name in an answer may search for it later. Watch English-language markets after launch, and use the October languages as a second, later comparison point if you sell in those markets.
2. Self-reported attribution. Add a free-text "How did you first hear about us?" field to demo and meeting booking forms. Code the answers monthly and tag mentions of Siri, iPhone, Apple or an AI assistant. Ask the same question on first sales calls. Our guide to AI search attribution covers how to connect these answers to pipeline.
3. AI prompt panels. Write a fixed set of 20 to 30 buyer questions. Each month, ask them on a supported iPhone set to English, run each question more than once, and record whether your company is named, how it is described, and which source links appear. Apple documents no API for this, so the work is manual. Run the same questions in ChatGPT and Google AI Mode for context, but report each surface on its own.

Verified Applebot requests in your logs are a useful leading check alongside these. Crawling is not citation, but a page Applebot never fetches cannot be used.

Google AI Overviews and AI Mode vs ChatGPT search vs Siri AI

The three surfaces rely on different crawlers and return very different data. Each cell below comes from the vendor's own documentation.

Google says its generative AI control covers AI Overviews, AI Mode and generative AI features in Discover, is not used as a ranking signal for other parts of Search, and does not affect training. Its performance report counts impressions, and GA4's AI Assistant channel excludes AI Overviews and AI Mode. OpenAI's crawler documentation says sites opted out of OAI-SearchBot will not be shown in ChatGPT search answers, and that robots.txt changes take about 24 hours to apply. For Siri AI, the Applebot page is the whole picture today.

Where to start this week

Steps 1, 2 and 6 are configuration checks, and step 6 has a hard date of September 15. Do those first, because they decide whether Siri AI can use your pages at all. Steps 3, 4 and 5 are policy and content work that helps across answer engines, not only Apple's. If you have not set a baseline yet, pull branded search data for the weeks before September 14 from Search Console now and start the prompt panel this month, so you have something to compare against. If you want a second pair of eyes on crawler access, snippet directives and entity consistency, request a free audit.