There is no shortage of advice on "ranking in AI Mode", and a lot of it describes ranking factors nobody outside Google can verify. I would rather start from what Google has actually published, because it is more specific than people assume, and then work out what it means for a B2B site that sells something complex to a small number of buyers. The short version is that Google has said, in its own documentation, that there is no special optimisation for its AI features, and it has described the one mechanism that matters most: query fan-out. Everything else in this piece is my reading of how those two facts apply to B2B content. Where I am inferring rather than quoting, I say so.
What Google has published about AI Mode
Google announced AI Mode's US rollout at I/O on 20 May 2025, in a post describing it as using a "query fan-out technique" that breaks a question into subtopics and issues many searches at once on the user's behalf. Its Search Central page, "AI features and your website", is the reference for site owners, last updated in December 2025 when I checked. It says that AI Overviews and AI Mode may use query fan-out, issuing multiple related searches across subtopics and data sources, and that while a response is generated the systems find additional supporting pages, which can surface a wider and more diverse set of links than a classic search. On eligibility, the page is direct: to be shown as a supporting link, a page must be indexed and eligible to appear in Google Search with a snippet, and there are no additional technical requirements. It also says site owners do not need new machine-readable files, AI-specific text files or special schema markup. Availability outside the US has been expanding since launch, so check Google's own announcements for your market rather than relying on a third-party summary.
What query fan-out means for B2B content
This is my inference, not Google's wording. If AI Mode turns a buyer's question into several narrower searches, then the pages that get cited are the ones that answer those narrower searches well, not only the page that targets the head term. Take a question like "how should a Series A SaaS company structure RevOps?". It plausibly fans out into sub-questions about which roles to hire first, which CRM to choose, what a reasonable tool budget is, and how RevOps differs from sales ops. A B2B site that has one broad service page will have one chance to be cited. A site that has a clear page or section for each of those sub-questions has several. This is not a new idea. It is the same topic coverage good B2B content programmes already aim for. What changes is the reward: being the best answer to a sub-question can now earn a citation in a response to a broader question you never targeted directly.
The fundamentals Google lists, applied to B2B
Google's best practices for its AI features are the normal Search ones, and each has a B2B-specific failure mode. Allow crawling in robots.txt and in your CDN or hosting setup: B2B sites behind aggressive bot protection sometimes block Googlebot without realising it. Make content easy to find through internal links: resource libraries that only surface content through on-site search or filters often leave pages orphaned. Keep important content in text: B2B sites frequently put their best material in gated PDFs, webinar recordings or slide embeds, which a search system cannot use as a cited answer. Provide a good page experience. Make sure structured data matches the visible text. And keep your Merchant Center and Business Profile information current where relevant. None of this is glamorous. In my audits, though, the most common reason a strong B2B site is missing from AI answers is plain technical: the content is gated, orphaned, or rendered in a way that is hard to index.
Content choices that make a page easier to cite
Again, this is practice rather than a published Google rule. Pages that state the answer early, in plain sentences, are easier for any summarising system to use than pages that build up to the point. For B2B that means leading with the direct answer, for example what a fractional CMO costs, how long a CRM migration takes or what a sales and marketing SLA should contain, then supporting it with specifics. Original material helps you stand out from the many near-identical pages on common topics: your own process, a template, a worked example, a calculator, or data you collected and can explain. Name your sources when you cite numbers, and avoid stating unverified figures as fact. Clear authorship and an about page that explains who is behind the advice support the "helpful, reliable, people-first" standard Google applies across Search. And write for the buyer's actual sub-questions, which you can find in sales call notes and the questions prospects send before a first meeting.
Controls and measurement
Google's page lists the controls that apply. Robots.txt rules for Googlebot govern crawling. The nosnippet, data-nosnippet, max-snippet and noindex directives limit what is shown from a page in Search, including in AI features. Google-Extended is a separate control that governs use for training and grounding in Google's other AI systems, rather than a Search control. Be careful with snippet restrictions on pages you want cited, since a page has to be eligible to show a snippet to appear as a supporting link. On measurement, Google says traffic from AI Overviews and AI Mode is included in the Search Console Performance report under the Web search type. In other words, it is not broken out separately there. To get a feel for AI visibility, track the queries and pages that matter to you over time, check landing pages for clicks that arrive from Google, and periodically search your core buyer questions in AI Mode yourself and note which sources are cited.
What I would ignore
Be sceptical of any list of "AI Mode ranking factors" with weights attached. Google has not published one, and the people selling them cannot see inside the system any more than you can. Be equally sceptical of advice to create special AI files or markup for Google's AI features: Google's documentation says it is not needed. That does not mean files like llms.txt are useless for other AI systems, just that they are not a Google AI Mode requirement. Avoid mass-producing thin pages for every possible sub-question. Google's spam policies on scaled content abuse still apply, and a hundred shallow pages are less likely to be cited than ten good ones. Finally, do not let AI visibility displace the basics of B2B demand. Being cited in an AI answer is valuable, but a buyer still needs a reason to click, a clear next step on the page, and a sales process that responds quickly when they do.
Sources
Google Search Central, AI features and your website: https://developers.google.com/search/docs/appearance/ai-features . Google, AI in Search: Going beyond information to intelligence (20 May 2025): https://blog.google/products/search/google-search-ai-mode-update/
FAQ
According to Google's Search Central documentation, no. Pages need to be indexed and eligible to appear in Search with a snippet, and Google says there are no additional technical requirements, no new machine-readable or AI text files, and no special schema markup needed for its AI features.
Google says sites appearing in AI Overviews and AI Mode are included in overall search traffic in the Search Console Performance report, under the Web search type. It is not reported as a separate line there, so you need to infer AI visibility from query and page trends plus manual checks.
It is Google's term for how AI Mode and AI Overviews may break a question into subtopics and run several related searches to build a response. For B2B sites, it suggests that answering a buyer's narrower sub-questions well can earn citations in responses to broader questions.