
- Google’s AI Mode now handles 75 million daily active users, and 93% of those sessions end without a single click to an external website.
- A major study of 680 million AI answers found that just 15 websites get 68% of all mentions inside ChatGPT, Perplexity, and Google AI. None of them are B2B IT and SaaS companies.
- YouTube citations in Google AI Mode surged from near zero to 40 million between January and May 2026. Video transcripts are now a GEO asset. LinkedIn posts are following the same.
- Google just made brand loyalty a formal input into Google’s ranking algorithm. Users who follow your site now see your content more often.
- AI-referred traffic converts at a much higher percentage versus organic search. Being mentioned in AI answers is not a nice-to-have. It is a revenue channel.
Your website traffic report is showing visits. Your CRM is showing fewer demos. The gap between the two keeps growing, and it is not because your content is bad. It is because search has quietly split into two separate systems, and most businesses are only visible in one of them.
The old system, the one your SEO strategy is built around, gives people a list of links and lets them choose. The new system, powered by ChatGPT, Google AI, Perplexity, Claude, and many more LLMs, reads everything and gives your audience a direct answer. If your business is not named in that answer, you simply do not exist at the moment the decision is being made.
This is Issue 001 of SEO Weekly from White Bunnie. Every Monday, we cover the most important changes in AI search and what they mean for B2B companies trying to grow through organic visibility. No jargon, No filler, just what changed, why it matters, and what to do about it.
This issue covers three signal-level events; Google’s AI Mode citation changes, the most comprehensive AI citation study ever published, and the YouTube GEO surge nobody is talking about yet.
For the full strategic picture behind everything we cover here, you may start with our complete guide to GEO and AI SEO for B2B SaaS companies.
Update 01: Google Changed How AI Search Decides Which Businesses to Show
Source: Google’s official blog, 6 May 2026

On 6 May, Google announced five structural changes to how AI Mode and AI Overviews display citations and links. In general, Google makes it easier for users to connect with ‘authentic voices’ across the web. The same week, Google removed the ‘Search’ button from Android devices and replaced it with ‘Ask Google’, the clearest signal yet that conversational AI is now Google’s primary interface. This matters because AI Overviews now appear in roughly one in four Google searches. For the kind of questions your buyers type ‘what is the best net-suite partner in new york’ or ‘how do I solve Y problem’, AI mode appears in what we estimate to be close to almost 40% of searches.
Here is what changed in plain terms. Google will now show links right next to the specific sentence they support, instead of lumping them at the bottom of the AI answer. It will highlight content from sources the reader already subscribes to or follows. It will pull in opinions and discussions from real communities; Reddit threads, forums, social posts; and name the person who said what. And on desktop, hovering over a link now shows a preview so people know where they are going before they click.
AI Mode already handles 75 million daily active users; a 4x increase since its May 2025 launch. Of those sessions, 93% end without a single click to an external website. Google’s own AI answers the question so completely that the majority of people never visit a single webpage. For businesses that built their growth strategy around blog content and how-to guides, which is very common for most B2B companies, this is the core reason of why website traffic is dropping in 2026.
What This AI Search Change Means for Your Business
Traffic from informational queries will continue to fall. That is the honest assessment. But this is not the same as visibility falling. Brands that are cited inside AI Mode answers are building shortlist presence at the point where B2B buyers are forming their vendor lists, before they ever run a branded search.
Being mentioned inside an AI answer is the new version appearing at the top of Google. The difference is that clicks are optional. The brand recognition happens regardless. And when that buyer is ready to make a decision, the name they already saw in the AI answer has a head start. The brands gaining ground are those whose content is structured to be extracted and attributed by AI citation engines.
If you want to know exactly how to structure content for AI search, ourtopical authority guide walks through the practical steps.
Action Plan for AI Search Update
Do This Week
- Read through your last three blog posts. Does each section open with a direct, clear answer in the first two sentences or does it build up slowly? Slow build-up means AI search skips it. Rewrite the opening of each section to lead with the answer, then explain it.
- Stop measuring only website visits. Add one new question to your monthly review: ‘Is our business being mentioned when someone searches our category on ChatGPT or Google AI or Perplexity or Claude?’ Search your top five keywords on these platforms and write down what comes up.
Do in the Next 60 Days
- Focus more structured content for the AI citation layer. This means structured, definition-first H2s, FAQ schema, and Article schema on every significant post. The technical investment is small. The citation returns compounds.
- Have a conversation with your team about what success looks like now. Traffic alone is no longer the right number to watch. Brand mentions in AI answers, direct enquiries, and demo requests are closer to the truth of whether your content is working.
- Brands that document citation-to-pipeline attribution now will have a measurable advantage in coming months.
Update 02: 680 Million AI Citations Were Analysed. 15 Domains Got 68% of Them
Source: 5WPR AI Platform Citation Source Index 2026, 1 May 2026, Largest AI citation study published to date
On 1 May, 5WPR released the AI Platform Citation Source Index 2026, the largest consolidated AI citation study to date. It synthesised 680 million individual citations across ChatGPT, Google AI Overviews, Perplexity, Gemini, and Claude. The findings are structurally uncomfortable for any brand that has been treating GEO as a blog formatting exercise.
Reddit is cited at approximately 40% frequency across all major LLMs. Wikipedia accounts for up to 48% of ChatGPT’s top-10 citation share. The top 15 domains capture 68% of all consolidated AI citation shares. The remaining 32% of mentions are shared among millions of other websites.
None of those 15 domains are specialist B2B publishers. The implication is direct; if your GEO strategy consists only of optimising your own website, you are competing for the remaining 32% of AI citation share while your competitors’ brands are being mentioned on Reddit, LinkedIn, and Forbes, which is where the AI answers are actually being built from.
There is also a warning buried in the data. ChatGPT’s reliance on Reddit dropped from roughly 60% to 10% in just six weeks in late 2025, because of a single change to how the AI was trained. Forbes and PR Newswire absorbed most of that shift. This tells you something important: where AI search gets its information can change quickly. Businesses that depend on only one source of visibility are one update away from disappearing.
What This Study Means for Your Business
Your organic ranking and your AI citation rate are increasingly independent variables. A page that ranks number one on Google does not automatically get cited by ChatGPT or Perplexity, particularly for professional and category queries. Only 11% of domains overlap between ChatGPT and Perplexity citation sources. Running a single GEO playbook across both platforms leaves the majority of AI citation opportunities unaddressed.
The businesses winning in AI search right now are not just the ones with the best websites. They are the ones whose name appears on G2, whose founder shares insights on LinkedIn, whose category gets discussed on Reddit, and whose results get covered in industry publications. Every one of those appearances is a signal AI search uses when deciding who to recommend.
If you want to understand where your brand stands in AI answers right now, the diagnostic is simple; search your primary category keywords on ChatGPT, Perplexity, and Google AI Mode. Document what comes up and what does not. Our topical authority guide covers how to structure content so your brand becomes a credible citation source.
Action Plan based on Study Outcome
Do This Week
- This week, go to ChatGPT and Perplexity and search the five questions your best clients were asking before they found you. Write down which companies come up. Write down which websites are mentioned. This is your competitive baseline.
- Identify the specific third-party platforms feeding AI citations in your category. Check whether your business has a G2 profile, a LinkedIn company page, and whether your category gets discussed on Reddit. These three platforms are among the most frequently cited sources for B2B software and IT services.
Do in the Next 60 Days
- Build a citation distribution strategy. Wikipedia entity presence, LinkedIn thought leadership content, PR Newswire distribution for research and data releases, and Reddit community engagement are all citation surfaces. Each requires a distinct approach.
- Run platform-specific GEO experiments. ChatGPT uses Bing’s index. Perplexity weights recency and community sources. Google AI Overviews pulls from organic top-10 results. Optimising for one platform will not produce cross-platform citation gains.
Update 03: YouTube Videos Are Now Feeding AI Search Answers, LinkedIn Posts Are Following Suit
Source: Ahrefs data / Lily Ray on X, corroborated by DesignRush SEO Roundup, 4 May 2026

YouTube citations in Google AI Mode were near zero at the start of January 2026. By early May, they had reached approximately 40 million. The data was surfaced by Lily Ray and corroborated by Ahrefs. YouTube already holds a 200X citation advantage over every other video platform in Google AI Overviews.
The mechanism is not complicated. YouTube automatically generates transcripts for every video. AI systems index and cite from those transcripts in exactly the same way they cite from blog posts. A 10-minute video generates roughly 1,500 – 2,000 words of citable transcript content. A B2B founder recording a 10-minute breakdown of their product category creates a citation asset that AI Mode can pull from.
The same pattern applies to LinkedIn. According to the 5WPR citation index, LinkedIn is the most-cited domain for professional queries across five separate AI platforms; viz. AI Overviews, AI Mode, ChatGPT, Copilot, and Perplexity. For B2B brands whose buyers are primarily professionals researching on these platforms, LinkedIn content is now a direct GEO input. Google also confirmed it is experimenting with Audio Overviews in Search Labs, and its VP of Search stated that video and audio content is actively being indexed and surfaced by LLMs.
What This Youtube and Linkedin Statement Means for Your Business
B2B SaaS brands that have built content strategies entirely on written blog content now have a measurable gap in their GEO coverage. YouTube and LinkedIn are citation surfaces that sit outside the standard content workflow.
Most B2B companies have built their content strategy entirely around written blog posts. That made sense two years ago. It leaves a measurable gap today. YouTube and LinkedIn are now confirmed active channels that feed AI search answers.
The businesses that start creating video content now; even simple, low-production recordings of a founder explaining their expertise, will be building an AI visibility asset before the channel becomes crowded. The same applies to LinkedIn. A consistent publishing habit under your own name, answering the questions your buyers actually ask, is now a direct input into whether AI search recommends your business.
Action Plan
Do This Week
- Audit your YouTube presence this week. If you have existing videos, verify they have accurate transcripts. If they do not, add them manually. A transcript is a citation surface. An uncaptioned video is not.
- Identify your top three category keywords and check whether any YouTube content currently appears in AI Mode or AI Overviews for those queries. This tells you whether video is already a citation factor in your category and who is winning it.
Do in the Next 60 Days
- Repurpose your highest-performing pillar posts as YouTube videos. The topic is already validated. The keyword signal is already proven. Recording a 10-15 minute walkthrough of the content creates a second citation surface from the same research investment.
- Treat LinkedIn content as a GEO channel, not a social channel. Structured posts that directly answer category questions and published consistently under a named author are regularly indexed and cited by AI platforms.
What to Watch in Coming Week
Google I/O; Google’s annual product announcement event is scheduled for 19 and 20 May. Based on everything that has happened in the last two weeks, expect announcements about how AI search is expanding, how more content types will be pulled into AI answers, and how Google plans to handle the question of giving websites more credit for the traffic they are losing. We will publish a response brief within 24 hours of the keynote.
If you want to know right now where your business stands in AI search, which keywords you are being mentioned for, which competitors are ahead of you, and what specifically needs to change, request a free AI Visibility Audit from White Bunnie. We send back a full breakdown within five working days. No commitment required.

