In today’s digital landscape, where artificial intelligence (AI) is transforming every aspect of our lives, search engines are no exception. A recent episode of Google’s Ads Decoded podcast shed light on key changes in search user behavior under the influence of AI. These changes directly impact content marketing strategies, especially mass video posting. According to data presented by Ginny Marvin, Google Ads Product Liaison, and Shivani Mohan, VP of Data and UX Research for Google Search, users are asking longer questions and starting their solution research earlier.
Transformation of Search Behavior under AI Influence
Users in AI search mode ask questions that are three times longer, actively use clarifying queries, and combine voice and visual search. This allows them to delve deeper into complex decisions, adding personal nuances to the search process. The processes of information discovery and option evaluation can now occur within a single dialogue.
Key Changes in Queries
Increased query length: Questions have become 3 times longer.
Multimodal search: More than 1 in 6 queries in AI mode in the US are multimodal.
Early research: Users begin their search at the problem consideration stage, not just for ready-made solutions.
“Users value the conciseness of AI summaries, but they also actively seek rich details, lived experiences, and authoritative opinions on the open web,” noted Shivani Mohan.
Content Strategies for Mass Posting in the New Search Era
For effective content distribution, especially with mass video posting, it is necessary to adapt strategies to changed user queries. The goal is to seamlessly integrate your content into their search journey, starting from the research stage.
Adapting Content to Search Trends
Users are looking for personal experiences and demonstrations that a product cannot provide in its description. Video reviews, podcasts, and forums are becoming key sources of information. Your mass posting service should take these trends into account, allowing you to quickly and effectively publish content that addresses such queries.
Identify needs: Analyze customer questions, sales team conversations, and search queries to identify recurring problems.
Test ad messages: Create ads that address identified needs, clearly explaining your offer.
Informative landing pages: Ensure that pages easily answer questions about price, compatibility, setup, and limitations.
Link to supporting content: Include links to relevant reviews, demonstrations, and customer success stories.
Conversion analysis: Track whether visits to comparative or educational pages lead to subsequent purchases. Evaluate results within the normal buying cycle.
Mass video upload: automation and reach
In an environment where users are looking for diverse and in-depth content, automatic publication of short videos becomes critically important. Use a single panel for TikTok, YouTube, Instagram and other platforms to ensure scheduled reach without manual labor.
Benefits of publishing automation
Time saving: While you sleep, videos are already being released on all platforms.
Scaling: Ability to publish 100+ videos per day to 10+ accounts.
Optimization: Using proxies and anti-detect for stable operation during mass upload.
Efficiency:Scheduled auto-posting allows you to plan content a month in advance, for example, a mass posting package for a month with a cost of 500 publications.
Frequently asked questions
How to choose a mass posting service, considering changes in search?
Choose a service that offers short video distribution to all key platforms (TikTok, YouTube Shorts, Instagram Reels) through a single panel. The ability to flexibly schedule and support mass uploads with proxies and anti-detect is important to ensure account stability and security. Make sure the service allows you to publish content that answers deep and multifaceted user queries.
How to adapt content for the early stages of the sales funnel?
Create videos that answer user questions at the problem research stage. These can be educational videos, comparisons, reviews, demonstrations of product use in various scenarios. Focus on solving problems and providing valuable information, not just direct promotion. Use video cross-posting for maximum audience reach on different platforms.
What is the role of the human factor in content when AI dominates?
Despite the rise of AI summaries, users still seek “live” experiences, reviews, and opinions from real people. Your content should include elements that demonstrate this experience: testimonials, user-generated content, expert opinions. This builds trust and helps users make decisions, which is especially important for automatic short video publishing, where quick information transfer is crucial.
Conclusion
Changes in search behavior caused by AI dictate new rules for content creation and distribution. Effective mass video posting today is not just automatic publishing, but a strategic approach to content distribution that responds to complex and multifaceted user queries. Integrate these principles into your strategy to ensure maximum reach and engagement. Start using tools for mass posting via API so your content works while you grow your business.
In a world where AI systems generate new queries at the speed of light, old hacking tricks of relevance are getting a second life. Greg Jarboe, a veteran SEO, recounts how back in 2003 he manually implemented what is now called a “fan query,” using the “Russian nesting dolls” principle. This method allows your content to be found for queries that don’t even exist yet, and is key to query repurposing in modern AI search.
Matryoshka SEO: How it worked before AI
Long before the terms “fan query” or “GEO” became common, Jarboe optimized press releases by embedding short phrases within longer ones. For example, if the target phrase was “flights to Philadelphia,” he would look for a four-word version, such as “cheap flights to Philadelphia,” and build the release around it.
This allowed for coverage of both short and long queries, making the content visible to a wide audience. This approach ensured maximum reach, even when search engines were less developed.
15% unknown queries: a Google constant
“I thought at some point most queries would have been done; people just ask the same thing over and over. But when we re-run those metrics, it’s always about 15%,” — John Mueller, Search Central Live NYC, March 2025.
Google first mentioned that 15% of daily queries are new in 2019 when introducing BERT. This figure remains stable despite the development of large language models. Multiply 15% by billions of Google queries, and you get hundreds of millions of unique formulations every day.
This is a huge layer of traffic that most SEO specialists underestimate. Understanding this phenomenon is critically important for effective query repurposing.
News as a query generator
Most of these new queries are not random. They are born from breaking news, new products, changed policies, or phrases coined by journalists. Press releases published on the day of an event are ideal for intercepting this traffic.
If the wording in the release accurately matches what a user types into Google, the content can rank for a query that did not exist at the time of writing. This demonstrates the power of timely content in capturing “new” search demand.
AI and reverse matryoshka: from broad to narrow
M.J. Cachon’s research showed that ChatGPT generates sub-queries that are also not random. Initially, the model uses general, conversational language, then narrows the search with the site: operator and begins to check exact quotes.
The use of quotes increases 25-fold between the first and last search. This is the “matryoshka” principle, but in reverse: AI starts broadly and narrows down to the exact phrase for verification. This AI approach to search opens up new opportunities for query repurposing.
Query Length: The New Reality of AI Search
According to Google data from May 2026, the average length of an AI query in the US is three times that of a traditional search query. Kachon’s data shows that one branded ChatGPT query branches into sub-queries, averaging seven words each.
Length is not a side effect of AI search, but a new norm that the industry has been moving towards for years. This means that content must be optimized for longer and more detailed queries to match AI behavior.
Practical Steps: How to Deceive Algorithms and Scale Content
To effectively use this knowledge for query repurposing, three key habits must be developed:
Look for nested phrases, not just keywords. For every three-word keyword phrase, find two or three four- and five-word phrases that naturally contain it. Build your introductory paragraphs or H2 headings around the longer version.
Publish at the speed of news. 15% of new queries are related to recent events. If your organization has a channel for same-day publication (press release, blog, rapid response page), this is your best chance to rank before competitors even know the phrase exists.
Write the verbatim answer as a quotable sentence. Kachon’s data shows that AI systems are increasingly verifying claims by looking for an exact quote from your content. Rewrite the sentence that answers the question so that it can be taken verbatim and still make sense.
These methods do not replace fundamental SEO principles but complement them, focusing on the long tail of queries that were previously underestimated. They allow for effective adaptation to changing AI search algorithms.
Frequently Asked Questions
What is a “query fan-out” in the context of AI?
“Query fan-out” is the process by which an initial user phrase or query branches into many more specific sub-queries generated by an AI system. This allows the AI to delve deeper into a topic and find more relevant information, narrowing the search from general to specific.
How is the “Russian matryoshka” principle related to SEO?
The “Russian nesting doll” principle in SEO means embedding short, high-frequency keywords into longer, low-frequency queries. For example, the phrase “airline tickets” can be nested into “cheap airline tickets to Philadelphia.” This allows content to be found for queries of various lengths, increasing its visibility and audience reach.
Why are 15% of Google queries always new?
These 15% of queries are always new because they are directly related to real-world developments: news, new product launches, changes in legislation, cultural phenomena. People search for information about what has just happened or appeared, generating unique formulations that did not exist before. This confirms the importance of timely content publication.
How do AI systems use quotes to verify information?
M.J. Kachon’s research showed that AI systems, such as ChatGPT, after a broad search, begin to verify information using exact quotes. They look for specific sentences in the content to ensure that the source actually states what the AI interprets. This emphasizes the need to formulate answers as precisely and quotably as possible.
Conclusion
In the era of AI dominance, understanding how systems generate and process queries becomes critically important. By applying the “matryoshka doll” principles, optimizing content for long and nested phrases, and publishing it promptly, you can outpace competitors and capture traffic that hasn’t even been conceived in users’ minds yet.
Start applying these hacking techniques today so that your content doesn’t just exist, but scales and tricks the matrix of search algorithms. Don’t miss the opportunity to become a leader in your niche by using advanced strategies of query repurposing!
Stop creating new content! It’s time to learn how to effectively use what you already have. Look into your team’s shared folders: there’s probably a dusty branded video that took six weeks and a five-figure budget to create, or a webinar recording with a minuscule number of views. Perhaps there’s a product landing page that has never been attached to an email by the sales department, or a transcript of a customer interview read only once. All this work wasn’t bad, but it stopped immediately after publication. In this article, we’ll explore how content repurposing can help you turn one video into 100 unique creatives.
Meanwhile, your quarterly plan demands even more: more materials, more channels, higher speed. You continue to create, the folder grows, and the true measure of team effectiveness is not the volume produced, but how much of it actually works. This gap costs real money, and it’s easy to overlook.
Video DNA: How to Break Down Content into “Molecules”
Blythe Morrow, VP of Product Marketing, directly addresses the math of channels: your buyer may be active on twenty channels, but “you can only fund and support content, communications, and the social aspect on five of them.” Everything you create for the other fifteen must spread further than just the “publish” button.
This issue surfaced in three recent conversations on the Supercharge Marketing podcast with experts who approach it from different angles:
Caroline Crawford, CEO and Founder of Cultiveight Communications, rebuilds content strategies for B2B teams.
Francisco Chamorro, Director of Marketing and Communications at BBSI, views content distribution as an operational task.
Blythe Morrow, an expert with seventeen years of experience in product marketing.
Plan the “Splitting” Before Creation
“They always talk about the ‘what,’ but not enough about the ‘why’ and ‘how it will be distributed,’” says Caroline, describing a typical content meeting.
The team discusses whether to make branded videos, ads, or blogs. But no one thinks about what will happen the next day. She has seen the end of this story many times: a company spends fifteen thousand dollars on branded videos, publishes them on social media, and, in her words, “tens of thousands of dollars go down the drain.”
The failure is not in the video, but in the fact that all the budget and energy went into its completion. “All the energy went into making it, and they think that in itself will bring success,” Caroline says. “But it doesn’t. It’s about how you apply it and how you maximize your investment.”
The solution is a planning stage that costs nothing and is almost always skipped: before starting production, think about how the finished material will be “taken apart.” One long article becomes the source for a lead magnet, a series of newsletters, a set of social media posts, and a video. One recorded conversation can provide material for weeks. Caroline describes this as “repackaging” content into “little pieces that tell your story in different ways.” Planning ahead allows you to shape what you film, write, and structure from the outset.
First Scale, Then Clone
Francisco comes to the same conclusion from the operational side. His advice to his past self: “if you’re going to build something, build it so it’s repeatable before you scale.” If you work too hard on one big launch, you get one launch. If you create a format your team can use again, the second time will be many times cheaper.
Repetition Matrix: No One Saw It the First Time
Most content is “retired” after one publication, based on the unspoken assumption that the audience has seen it and moved on. Francisco believes this assumption is flawed. “If you spend time and value this material, know that it won’t be seen 100% the first time,” he says. “Therefore, it’s important to reschedule it.”
“Your job is to give content the longest possible life support,” — Francisco Chamorro.
This means that cadence (publication frequency) is part of the asset, not just a nice-to-have. Francisco applies this same concept internally and externally, based on two principles: consistency and relevance. “People love repetition.” Communications go out on a predictable day, so the audience forms expectations. And the calendar is edited by data, not opinions: “When we create a topic that works well, we do more of that. When a topic doesn’t work, we discard it.”
Caroline adds that this doesn’t turn into spam if the content is created with intent. Many B2B teams are so afraid of being “noise” that they publish every two months, draining their funnel. She argues that noise is a function of sameness, not frequency. Repeat the same message in the same format, and people will tune out. Repeat the same message, but “repackaged,” and it will feel new. That’s why the planning stage mentioned earlier is so important. A team that has planned six formats in advance has six legitimate ways to repeat the same thing.
Expert Approach: Creating “Raw” Material
In the Room with Clients: Capture Live Video-DNA
Blythe’s answer to “what to do next” is to stop guessing. Product marketers, she says, “get into their own echo chamber,” and then “play broken telephone with sales,” where sales gives feedback to customers, and marketing gives it to product, and the original words are lost by the second stage.
Her recipe: get out of the office, talk to sales, sit in on customer calls. Instead of webinars, which she describes as “one-to-many lectures,” she runs small workshops. About twelve people. Cameras on. An honest conversation about what’s happening in their industry, not a product presentation. Product managers and customer success are present and listening.
Running these meetings regularly yields quick results: 150-160 people a year with whom you’ve had a real conversation, who then populate your LinkedIn and Slack communities and amplify your message publicly. Each sales manager is assigned one roundtable, and the follow-up call becomes “warm” because, as Blythe says, “they are much more likely to pick up the phone and talk to you because you’ve already shown that you value them.”
And the content problem solves itself as a side effect. Each workshop, she says, “creates incredible content that you can then give to the demand generation team to slice and dice 100 different ways and use for the many campaigns we’re trying to run at the same time.”
The raw material, created in the language of real customers, is the one that will be used because it already sounds like the market.
Internal Audit: If Sales Doesn’t Use It, It’s Not Content, It’s a Warehouse
Francisco is direct about priorities: “I don’t care how cool your video or campaign is, your internal users have to believe in it first. If not, it won’t be successful.” Beautifully crafted material without internal adoption has an audience of zero, so “spending the time and effort to make sure what you’re doing helps your sales team is critical.”
Blythe agrees and clarifies: product marketing is “the custodian of the message, the words we use to describe what we sell.” And the artifact that makes this real is a detailed, open-to-use messaging guide. Her guide covers long and short-form positioning, industry tailwinds and headwinds, every persona (where they live, what they struggle with), and the same depth on competitors. Then the part most teams miss: “keep it updated, share it often, refer to it constantly, and keep adding to that library.”
She also names a failure mode that’s worse than a vague message. Confused copy at least announces itself. “When it gets specific but in the wrong way because you haven’t thought about who you’re selling to, then that becomes dangerous,” she says, “because you don’t know if it’s working or not.” This is the asset that gets created, approved, published, and quietly underperforms for a year while everyone assumes the channel is the problem.
Reverse-Engineering the Path: From Clip to Conversion
Content that never converts isn’t always left unseen. Sometimes it’s seen, then abandoned because the reader has nowhere to go. Caroline’s version of the question: you’ve published repurposed clips, so “what’s the path back to you?” If you haven’t built a route from the public channel to your website, trial, or sales team, “then you’re making it up as you go,” and that’s where the cracks appear.
She’s equally direct about the order in which teams work. “I think people do it backwards,” she says. “They say, ‘Oh, let’s make this lead magnet, and then we’ll figure out the email.’” Flip that. Decide what the email program is for, then build the lead magnet that feeds it. Her core principle applies to every step of the journey: “the more thought effort it takes for them to do what you want, the less they’re going to do what you want.”
The same discipline applies to your tech stack, and this is where Francisco is most helpful. He sees two traps in almost every team: “One is that they let the tools drive their strategy. And the other is that they don’t use their tools to their full potential.” Before buying anything, he offers one question: “if you’re paying for software, make sure to ask yourself, am I using it to its full potential?” And about the integrations everyone builds because they can, his rule is worth writing down: “integrating for the sake of integrating is not something I would spend my efforts on.”
Blythe would add that none of this is cause to be ashamed of a lack of resources. “Everybody is under-resourced,” she says. The teams that get ahead aren’t the ones with more people, but the ones whose existing work covers more of the channels where their buyers are.
The Bottom Line: How to Cheat the Algorithms and Scale Content
Don’t start by planning a new asset. Start by auditing last quarter. For each piece of content, ask three questions:
Was it ever republished after the first week?
Was it ever used by sales?
Was it ever cut into a second format?
Anything that scored zero on all three isn’t a content problem. It’s a distribution and adoption problem, and creating new won’t solve it. Then, pick one asset that deserved better, and give it the second life it never got.
Turning one asset into many is exactly what Lumen5 is built for. Teams use it to transform articles, transcripts, and documents they already have into polished video, with brand kits that keep colors, fonts, and logos consistent no matter who on the team is making it. That means less time rebuilding from scratch and more control over how your brand shows up on every channel.
Ready to dive deeper into what full-funnel B2B marketing looks like at scale? Listen to the full conversations on the Supercharge Marketing podcast, available everywhere you listen to podcasts.
Lumen5 helps B2B marketing teams produce more video content faster and at a fraction of traditional production costs. Learn more or book a demo.
Frequently Asked Questions
What is content repurposing?
Content repurposing is the process of re-adapting existing content into new formats or for new audiences to maximize its value and reach. For example, a webinar can be turned into articles, short videos for social media, podcasts, and infographics.
How to make 100 unique videos from one?
To create many unique videos from a single source, use cutting, reformatting, and adding new elements techniques. For example, from a long webinar, you can cut dozens of short clips for TikTok, Shorts, or Reels, changing the background, adding subtitles, voiceovers, or musical accompaniment. Tools like Lumen5 automate this process.
Does repurposing help bypass ad bans?
Yes, content uniqueization through repurposing can help bypass ad bans. Publishing the same creative on different platforms or repeatedly using the same “clone” increases the risk of a ban. Creating variations with altered “video DNA” (visual elements, voice timbre, structure) helps anti-fraud algorithms perceive them as new content.
Should I order new content or repurpose old content?
It is optimal to combine both approaches. Creating new, high-quality “raw” material, specifically designed for subsequent “splitting” into many unique units, is the most effective strategy. This allows for significant budget and time savings compared to constantly creating content from scratch.
How does repurposing affect SEO?
Content repurposing positively affects SEO because it allows you to create more relevant content on key topics, increase the number of internal and external links, improve behavioral factors (due to format diversity), and enhance visibility in search engines. It also helps reach different audience segments who prefer various information consumption formats.
The era of fragmented queries is fading. Data from a year of Google’s AI search mode shows: users are asking complete questions, not fragments of phrases. This is critically important for stream highlights and any other content: now the main information should be at the beginning so that AI algorithms can instantly extract it and present it in their reviews. If you bury the essence in long introductions, your content risks going unnoticed.
Evolution of Search: From Keywords to Dialogue
Since the launch of AI mode, user queries have fundamentally changed. According to Shivani Mohan, Google’s VP of Data Science, AI mode has reached 1 billion active users per month. But it’s not just about the increase in the number of queries.
Query length has tripled: The average query in AI mode in the US is three times longer than traditional queries.
Multimodal search: More than 1 in 6 queries now include images, voice input, or interactive dialogue.
Growth of images: Image-based queries are growing by 40% monthly.
Follow-up questions: Users ask clarifying questions, indicating the depth of their search.
The most common first words in AI mode queries are “what,” “how,” “I,” “is,” and “can.” This demonstrates a shift from searching by individual words to task- and question-oriented search, similar to communicating with a human.
Five User Behavior Modes in AI Search
Mohan identifies five main user behavior modes, each requiring a specific approach to content:
Explore: Open-ended queries for brainstorming. Growing 30% faster than overall AI mode traffic.
Decide: Queries with comparison words (“which of,” “which one”). Growing 40% faster.
Learn: Understanding new concepts and professional development.
Create: Queries for image creation have tripled since the beginning of the year.
Do: Planning (workouts, travel, budgets). Growing 80% faster in the last six months.
“This is not just a change in ranking factors; it’s a transformation of user behavior. If you’re still building content around keywords, that model is rapidly becoming obsolete,” says Greg Jarboe.
To make your content effective in the new reality of AI search, follow these five steps:
1. Start with the main point
Place the main definition, key metrics, or main conclusions directly in the first sentence of each main section. This allows AI algorithms to quickly extract the necessary information.
2. Use specifics
Formulate sentences with specific brand names, geographical markers, exact dates, and verified numerical values. Avoid vague generalizations so that your text is “mathematically readable” for search engines and reduces the risk of AI “hallucinations.”
3. Structure for scanning
Organize the main text into short paragraphs (2-3 sentences). After section headings, use brief summaries, ordered lists, or structured tables. This facilitates segment extraction for AI and ensures readability for humans. This is especially important for mass cutting of 100+ clips, where each fragment must be self-contained and informative.
4. Prepare answers to follow-up questions
If multi-turn dialogues are growing by 40% per month, your content should answer the second and third questions a reader will ask after the first. Structure long content so that each section can serve as an answer to a specific follow-up question. This increases the likelihood that this particular “chunk” will be provided by the AI mode. For cutting game streams, this means creating clips that answer specific viewer questions.
5. Optimize multimodal content
With the growth of image queries (40% monthly) and a threefold increase in image creation requests, alt-text, image context, and visual content quality are becoming key ranking factors. They are no longer secondary but are part of what the AI mode reads.
Conclusion: From “assembly line” to “smart content production”
Success in modern SEO is not about flooding the internet with templated texts, but about mastering structural clarity that provides immediate and verified benefits to both algorithms and readers. Just as 19th-century journalists adapted to the telegraph, we must rethink our approach to content creation for the AI era. Cutting streams without losing quality, creating highlights, and cropping 9:16 from horizontal — all this requires not just technical skills, but also a strategic approach to information delivery. Remember that the person who cuts streams must now think like an SEO specialist so that every Twitch clipping or podcast video cutting works as effectively as possible to attract an audience. Contact us to find out how we can help you with this.
Frequently Asked Questions
How long does it take to cut a 2-hour stream?
The cutting time depends on the complexity of the content and the desired number of highlights. On average, professional cutting without loss of quality of a 2-hour stream into 5-10 clips can take from 1 to 3 hours, including adding subtitles for cuts and optimization. If you need bulk cutting of 100+ clips, automated tools such as Opus Clip analog service are used, which significantly speed up the process but require subsequent manual refinement.
What moments are usually cut from streams for highlights?
For highlight editing, the most emotional moments, key game events (e.g., “kills”, “win rates”), funny jokes, important announcements, Q&A sessions, and fragments that evoke a strong audience reaction are usually cut. The goal is to create cuts for TikTok from Twitch or YouTube that quickly grab attention.
Are subtitles needed for stream cuts?
Yes, subtitles for cuts are extremely important. They improve content accessibility, allow videos to be watched without sound (e.g., in public places), and increase engagement. In addition, subtitles contribute to better indexing of content by search engines, which is important for stream monetization through cuts.
What is the price for 1 hour of stream for cutting 20 clips?
The price for cutting 20 clips from a stream for 1 hour can vary depending on the specialist’s qualifications, the complexity of the editing, and the need to add graphics and subtitles. Approximately, the cost can range from 1000 to 3000 rubles per hour of source material, if it is a high-quality work taking into account SEO optimization and 9:16 cropping from horizontal for different platforms.
In August, SEO specialist Anastasia Kuru noticed queries in the Search Console report that didn’t look like search queries: “Yes,” “Yes, continue,” “Yes, price.” John Mueller confirmed that Search Console includes data from AI Overviews and AI Mode in the overall report, and subsequent questions within AI Mode are recorded as new queries. This opens up the opportunity to analyze fragments of AI dialogues that appear in your statistics.
In this article, we’ll break down how to recognize such fragments in Search Console data, how they differ from regular queries, and how to use this information for optimizing content for AI search.
Seven types of AI fragments in Search Console data
I analyzed 16 months of data from my site and identified seven categories of such queries. Each has its own origin and characteristics.
Short answers
These are single words or phrases like “yes,” “of course,” “show me.” They occur when a user responds to AI during a dialogue, and the response is treated as a search query. Positions in such cases are taken from the answer block, not from the search results.
Comparisons in dialogue
Phrases like “what about Resend?” or “what if we try Gemini?” — the user asks AI to compare an alternative with the already received answer. This signals a specific interest in your content.
Questions addressed to a conversational partner
The grammar of such queries implies the presence of a listener: “Can Meta Ray-Bans be hacked?”, “How do I sell this?”. They differ from regular search phrases.
Synthetic prompts from SEO tools
Regular queries, for example, “evaluate [company] by [criterion]” or “. my location is usa.” They repeat daily for months, indicating automation.
Full instructions for agents
Machines sometimes log entire prompts, for example, “find on the web… return 3 most relevant results… don’t invent URLs.” Such strings end up in the report as queries.
Errors and table headers
People and pipelines search as-is. In my data, there was an entire row from a ranking tracker CSV file that received 146 impressions.
Long queries without markers
Phrases of 10+ words without obvious signs of dialogue. The classifier sends them for manual review, as they could be either AI fragments or regular long-tail queries.
How to distinguish AI fragments from regular queries
The key difference is the addressee. A regular long query is addressed to no one, while a conversational one is directed at a conversational partner. Four signals reveal the conversational nature:
presence of an address;
interrogative construction;
incomplete sentences;
conversational particles.
Length itself is a weak signal, so queries of 10+ words with interrogative intonation are classified as conversational, while those without markers are sent to quarantine.
Practical conclusions from the analysis
On my site, AI fragments began appearing in March 2026, steadily 20–30 impressions per month. For example, the query “yes” gave 110 impressions and 6 clicks at an average position of 4.5 — this is impossible in regular search results, but typical for an AI answer block.
Comparative queries perfectly correlate with my articles: “what about Claude?” led to the WebMCP guide, “what about Xcode?” — to the post about Xcode. This shows that users actively compare alternatives, and your content can be mentioned in AI answers.
Frequently asked questions
Can AI Overviews be separated from AI Mode in the data?
No, Google combines them in the report as web search. The fragment indicates the fact of a dialogue, but not the specific type of surface.
Why does Google hide queries and clicks in the Generative AI report?
The report shows impressions, pages, countries, devices, and dates, but not queries and clicks. The API does not support this data, and the BigQuery export does not contain an AI column. The only way to get data is the export button in the interface.
How to set up a classifier for your site?
The classifier genai_conversation_queries is available in the free Search Console MCP v2.4.0. Install MCP with one command, select your project, and run the analysis. For large sites, use the BigQuery version v4.1.0.
Conclusion
Analyzing AI fragments in Search Console provides unique insights into how users interact with AI and find your content. Use the classifier to uncover hidden opportunities and optimize content for AI dialogues.
Start with the free tool and get data that most SEO specialists miss.