On October 15, 2026, Google shipped a set of updates across AI Mode and AI Overviews — and one of the five changes is structurally different from the rest. Google is testing a dedicated section at the top of an AI response that surfaces “one or two official places on the web” for the task the user is trying to complete, each with context about why the link is useful. Everything else in the update — more inline links, prettier carousels, clickable creator profiles — is incremental. This one changes what a citation is for: from “a source the answer drew on” to “the place the user should go.” If your B2B SaaS site can be the official source for a task your buyers actually do, that section is the most valuable placement in Google’s AI surface right now. This is what changed, what the data says about the stakes, and the practical work of becoming that official source.
What shipped on October 15
All five items below are from Google’s own announcement (Hema Budaraju, VP Product Management, October 15, 2026: https://blog.google/products-and-platforms/products/search/updates-ai-mode-overviews-explore-web/). Being precise about what is live versus tested matters here, because the announcement mixes both:
- More inline links, higher up. Google “made continued and substantial increases to the presence of inline links” in AI experiences and is showing them “higher up in responses more frequently.” Inline links are the links embedded in the words of the answer itself — the ones users click without leaving the response.
- A new top-of-response “official” section (testing). “Sometimes, there are only one or two official places on the web to complete a specific task or get the exact info you need. We’ll be testing a new display in Search AI experiences to prominently surface these links in a dedicated section at the top of a response, with context about why each link is useful.”
- News and Preferred Sources carousels, wider rollout. The existing carousels for timely articles are expanding “to additional locations and languages globally” and now “highlight your Preferred Sources.”
- More prominent end-of-response links. Related-article links at the end of some responses get image thumbnails and “a more clickable design.”
- Clickable creator profiles (testing). Creator and publisher names on video results in AI Mode link to their Search profiles.
One supporting number from the same post: usage of Google’s Preferred Sources feature “has grown by more than 7x since the start of the year.” Google is clearly investing in the mechanism where users declare which sources they trust — which makes it plausible that “official source” judgments in AI answers will matter to Google, not just to users.
What the post does not include, and you should not expect: any documented selection criteria for the new top section, a rollout timeline, or a query list. It is a test, and Google says the updates will be “refined… based on what works.” Anything more specific you read about how the section is decided is speculation, including from us.
Why the stakes are now, not later
The new section only matters if AI answers are actually showing up on the queries that matter to you. They are — increasingly on branded queries, which changes the risk profile of being the un-cited brand:
Ahrefs analyzed 232.8 million US desktop Google search-result crawls from July to September 2026 and found AI Overviews appeared in 82.91% of US desktop results for branded keywords as of September 29, 2026, up from 61.3% in July and 65.1% in August (Ahrefs, Ryan Law, October 2, 2026: https://ahrefs.com/blog/ai-overviews-on-branded-searches/). In a separate panel of 100 global brand names checked weekly over the same period, 76.2% of the branded AI Overviews appeared in position one — meaning the Overview, not the brand’s own website, sat at the top of the brand’s own query.
The second half of that study is the uncomfortable part for brand control. Where branded AI Overviews cited external sources — sites other than the brand queried — the most common were Wikipedia (71.9% of AIOs citing it), YouTube (38.6%) and LinkedIn (22.8%). Same Ahrefs study; same vendor-pipeline caveats as any single-vendor crawl data — the direction is the point, not the decimals.
Read those two facts together with the October 15 update: AI answers are already dominating branded queries, the external sources currently shaping brand narratives are ones you do not control, and Google is now building a dedicated slot for “the official place” on top of the response. If that slot is empty of you, the default for “official” keeps drifting toward Wikipedia and the rest.
How to be the “official” source (four workstreams)
There is no published formula for the new section, so this is judgment, not a recipe. The pattern Google describes — “one or two official places to complete a specific task” — maps onto things a B2B SaaS site either already has or lacks:
1. Entity clarity: one unambiguous answer to “what is this company, and where does it live?” AI systems and Google’s entity graphs decide “the official site” from consistent signals — the same name, same logo, same description, same ownership, across your site, your social profiles, your docs, and third-party references. If an AI engine has to reconcile five versions of your company name and two domains before it can decide what your brand is, you will not win a slot that is explicitly about the official place. This is the work we break down in entity clarity for AI, and it is the longest-lead item in this list — do it first.
2. A brand page that reads as the home for the task. The new section surfaces pages “with context about why each link is useful.” A page that is a marketing hero with no task language is weak evidence. The pages that fit the slot are the ones where a user lands and completes something: a product page that names the task it performs, a login or dashboard entry point, a status page, a calculator or tool. Check what your top-of-brand-query SERP actually contains and ask which of your pages would survive as “the place to do [X].”
3. Documentation as an official surface. “Complete a specific task” is often a developer or operations phrase. If your buyers configure integrations, parse webhooks, or run API calls, your docs domain is arguably the official place — and it is usually the best-structured, most answer-shaped content you have. Make sure docs are crawlable, consistently named, and discoverable from the main site so the two surfaces read as one entity.
4. Third-party corroboration, earned the legitimate way. Ahrefs’ branded-query data shows Wikipedia and LinkedIn currently doing much of the “official” work for brands. For a B2B SaaS, the corroboration set is smaller and more winnable: an accurate Wikipedia or Wikidata entry where you genuinely qualify, a clean LinkedIn company page, consistent listings in directories your category actually uses, and review-platform presence. None of this is “build authority” fluff — each is a signal that an independent source confirms what your own site claims, which is exactly what an “official source” judgment requires.
How to check your position now
You cannot request entry into the new section, and you may not see it in your market yet — it is a test. What you can do, this week:
- Read your GSC generative AI report. The report (rolled out globally August 31, 2026) shows which of your pages Google’s AI features actually feature — your exposure map; archive this month’s numbers as a baseline artifact, since the rollout will make month-over-month deltas the real signal.
- Run your brand query and five task queries in AI Mode and AI Overviews by hand. Note whether the answer links to you, what it links to instead, and whether the answer’s “official” framing (login, pricing, setup) points at your site. Ten queries, thirty minutes. Our cross-surface walkthrough is in which AI surfaces actually cite B2B SaaS brands.
- Compare with last quarter. If your brand query’s AI answer cited a Wikipedia page or a YouTube channel last month and cites your site this month, the October 15 rollout is already working on your category. Record both — absence in the top section is a finding, not a non-event.
The honest summary: the October 15 update does not reward content volume, and no one — including Google — has published what earns the new top slot. What it does do is make the official-source question visible and persistent in AI answers, and it does that on a surface (branded queries) where the current answer is often not you. Entity clarity, task-shaped pages, clean docs, and corroboration are the levers you can actually pull on that question. The rest is measuring whether it is working, monthly, on your own queries — which is exactly what the methodology page covers.