Key Takeaways
AI tools like ChatGPT, Perplexity, and Google AI Overviews now influence nearly half of all local business searches, yet only 1.2 percent of local businesses are being named in AI responses. Home-service contractors who post structured YouTube videos, publish transcripts as on-page text, and embed videos on relevant service pages are earning citations in as little as four days, with no ad spend required.
- 45 percent of consumers used AI to find a local business as of February 2026, up from 5-6 percent the year before, according to BrightLocal.
- Google Maps rankings do not transfer to AI recommendations. Fewer than half of businesses that rank well in Maps get named by AI engines.
- YouTube is cited 21 percent of the time in AI answers across ChatGPT, Perplexity, and Google AI, ahead of Reddit, Wikipedia, and every other source, per an Ahrefs study of 3 million searches.
- AI reads video transcripts, not footage. Every city name, service type, and price range spoken on camera becomes quotable text for AI engines.
- Videos of 8-15 minutes earn 94 percent of AI citations. Short-form content drives channel views but rarely earns citations.
- 41 percent of AI-cited YouTube videos have fewer than 1,000 views, meaning channel size does not determine citation frequency.
- One Bear Restoration video campaign produced 24 AI mentions out of 142 tested queries, with zero paid ads and no SEO agency involvement.
Is AI Recommending My Business
Most local home-service contractors are not being recommended by AI. Data from PushLeads’ own PushRank live tracker, which logged 2,794 real queries across 42 clients and 813 unique visibility scenarios, shows YouTube citations appearing 536 times more frequently than any other source in Google AI Overviews and Perplexity. ChatGPT named a local business in only 1.2 percent of applicable queries. That figure comes from tracking real customer questions, not laboratory simulations.
The gap between Google search visibility and AI visibility is the central problem contractors face right now. A business can hold the top position in Google Maps and still be completely absent from AI-generated answers, because AI engines pull from different source pools. Google Maps rankings are built on reviews, proximity, and category signals. AI citations are built on authoritative written and spoken content that directly answers a specific question a potential customer would ask.
Checking whether AI is recommending a business requires testing real buyer questions across multiple platforms. Searching “who handles water damage restoration in Asheville NC” in ChatGPT, Perplexity, and Google AI Overviews on the same day will show whether a company’s name appears. If it does not, the business is not part of the new discovery layer that now accounts for 45 out of every 100 local business searches. The sections below explain what drives those citations and how to pursue them systematically.
Want more customers from Google & AI search?
Get a free SEO audit of your site — see exactly what to fix first.
How Much the AI Search Shift Has Already Changed Local Discovery
The share of consumers using AI to find a local business moved from roughly 5-6 percent in 2024 to 45 percent in early 2026, according to BrightLocal’s February 2026 survey. That is not gradual adoption. It is a structural change in how people locate service providers in their area.
Google remains the most-used discovery channel, but its share dropped from 83 percent to 71 percent in a single year. That 12-point loss did not go to Bing or Yelp. It went to AI tools. As of February 2026, 31 percent of people used ChatGPT and 23 percent used Google’s AI mode to find local businesses. These are not tech-early-adopters. They are ordinary homeowners deciding who to call when a pipe bursts or a roof starts leaking.
Why does Google’s market share drop matter for a local contractor?
A 12-point drop in Google’s share means roughly 1 in 8 searches that previously went through Google’s traditional results are now going through AI interfaces. For a contractor receiving 40 leads per month from Google, that shift could represent 5 or more leads per month that bypass Google entirely. Over a year, that compounds into a meaningful revenue gap for businesses that have not adapted their content strategy.
Is the 45 percent AI adoption figure specific to any one type of consumer?
The BrightLocal data covers general local business searches and is not restricted to a single age group or geography. Home-service categories including HVAC, plumbing, roofing, and water damage restoration appear in the survey’s scope. Contractors in mid-size markets like Asheville, Fayetteville, and Knoxville face the same adoption curve as those in larger metros, because AI tools are available nationally at no cost to the user.
Does this mean Google optimization is no longer useful?
Google optimization remains useful because Google still holds a 71 percent share and because Google’s own AI Overviews are now one of the primary citation surfaces. However, optimizing only for traditional Google Maps rankings is no longer a complete strategy. A contractor needs to address both the Maps game and the AI citation game with different tactics, because the ranking signals that govern each are distinct. Solid local SEO still forms the foundation, but it now needs to be layered with content that AI engines can quote directly.
Why Google Maps Rankings Do Not Carry Over to AI Results
Fewer than half of businesses that rank well in Google Maps appear when someone asks an AI tool for a local recommendation. The two systems use fundamentally different inputs. Google Maps rankings depend on proximity, review volume and rating, category accuracy, and engagement signals from Google Business Profile. AI recommendation systems depend on whether a business has produced quotable written or spoken content that answers specific questions.
The Pew Research Center analyzed 68,000 real searches and found that when AI answers a query at the top of a results page, click-through rates drop to roughly 1 percent. For traditional Google results without an AI answer, click-through is around 8 percent. That means the informational layer of local discovery has moved away from the click and toward the direct answer, and that direct answer pulls from sources that Maps-focused businesses have rarely developed.
What sources does AI actually pull from when answering a local service question?
An Ahrefs study published in June 2026 analyzed 3 million searches and found YouTube cited 21 percent of the time in AI-generated answers. Reddit followed at around 8 percent. Wikipedia came in at approximately 6 percent. All other sources combined represented the remainder. Across ChatGPT, Perplexity, and Google AI, YouTube is cited roughly 200 times more frequently than any other video platform, according to BrightEdge data drawn from 30 million sources.
Why does YouTube outperform every other platform by such a wide margin?
AI engines do not watch video. They read the transcript that accompanies video content. YouTube’s transcript infrastructure is mature, searchable, and indexed by every major AI engine. When a contractor speaks on camera and says “water damage restoration in Asheville typically costs between $1,500 and $5,000 depending on square footage and moisture depth,” that sentence becomes a citable text passage. Vimeo, TikTok, and other video platforms receive 0.1 percent of video citations in the BrightEdge data, because their transcript infrastructure is not as consistently indexed.
Does a high Google Maps ranking help at all with AI citations?
Indirectly, yes. A strong Maps presence supports the overall authority profile of a business, and some AI engines do weight authority signals when deciding which local businesses to name. However, the direct citation mechanism for AI answers runs through YouTube transcripts, on-page written content, and third-party mentions, not through Maps position. A business can rank number one in Maps and generate zero AI citations if it has no video or written content that answers specific buyer questions.
What PushLeads’ Own Tracking Data Shows About AI Citation Patterns
PushLeads operates a live tracking system called PushRank that monitors AI citation behavior across ChatGPT, Perplexity, and Google AI Overviews using real client data. From June through September 2026, the tracker processed 2,794 real questions drawn from 42 client accounts across multiple home-service industries, generating 813 unique question-and-visibility scenarios.
The raw citation count from that tracking period shows YouTube appearing 536 times more frequently as a cited source than any other source type. Within the YouTube citation data, Google AI Overviews produced 421 citations and Perplexity produced 115. ChatGPT produced zero YouTube citations during the same period, which is consistent with ChatGPT’s known reliance on text-based sources like Reddit threads and published articles rather than video transcripts.
What does the Bear Restoration case show about citation frequency for a small business?
Bear Restoration is a restoration company in Northwest Arkansas. It is not a large franchise or a nationally recognized brand. PushRank tracked 142 real buyer queries of the type a homeowner would ask about water damage or restoration services. Bear Restoration was named in 24 of those 142 answers. That is one named citation for every six queries. The breakdown was 17 citations from Perplexity, 4 from Google AI Overviews, and 3 from ChatGPT. The company ran zero paid ads during the tracking period and was not under contract with an SEO agency at the time.
How does ChatGPT’s citation behavior differ from Google AI and Perplexity?
ChatGPT did not cite YouTube videos at all in PushLeads’ live tracking data across the June to September 2026 period. Google AI Overviews and Perplexity both cited YouTube frequently. For contractors who want visibility specifically in ChatGPT responses, the channel that appears most effective in the tracking data is Reddit. Active, detailed participation in relevant Reddit threads, particularly in local and home-improvement subreddits, creates the type of text-based source that ChatGPT draws from when constructing local recommendations.
Is citation frequency consistent across industries, or does it vary by trade?
The PushRank data covered 42 clients across multiple industries, so variation exists. Trades with higher consumer curiosity and higher average ticket prices, such as restoration, HVAC, and roofing, tend to generate more AI queries because consumers are more likely to research before calling. Trades with lower average tickets and faster decision cycles, such as locksmith or drain cleaning, generate fewer AI queries because consumers often call the first name they see. Citation opportunity scales with the research intensity of the buying decision. Understanding where a specific trade falls on that spectrum helps prioritize content investment.
How Video Transcripts Become AI Citations
AI engines read transcripts, not footage. A video that shows a technician replacing a water heater without any spoken explanation generates no text for an AI to quote. A video where the technician says “water heater replacement in Asheville typically runs between $900 and $1,400 for a 40-gallon tank-style unit, not counting permit fees” gives the AI a specific, attributable sentence it can surface when a homeowner asks that question.
The practical implication is that silence, B-roll narration that lacks specifics, and graphic-only content all produce zero citation value regardless of production quality. An agency-produced video with motion graphics and a professional voiceover that never states a city name, a service type, and a price range will earn fewer AI citations than a phone-recorded video where the contractor speaks for eight minutes answering one specific question in conversational language.
What should a contractor actually say on camera to earn citations?
The four categories of spoken content that appear most frequently in AI citations from home-service videos are: the city and neighborhood the business serves, stated explicitly rather than implied; the specific service being discussed, named by its common buyer term rather than technical jargon; a price range or cost estimate with context for what affects the number; and a direct answer to a question a homeowner would type into a search bar. Saying “roof repair in Asheville” and following it with “most shingle repairs for storm damage run between $300 and $900 depending on square footage and pitch” creates two citable sentences in one breath.
How long should a video be to maximize citation probability?
94 percent of YouTube videos cited in AI answers are 8-15 minutes long, not short-form content under 60 seconds. Short videos may drive subscribers and views, but the transcript is too brief to contain enough specific, answer-dense language to compete for citations. An 8-15 minute video on a single question produces enough transcript text for an AI engine to locate and surface a relevant passage. Posting one well-structured video in that length range per week will produce more citation opportunities over a quarter than posting daily 30-second clips.
Does video view count affect whether AI cites it?
41 percent of YouTube videos cited in AI answers have fewer than 1,000 views. Channel size and subscriber count are not the deciding variables. The deciding variable is whether the transcript contains a specific, accurate, locally relevant answer to a question the AI is trying to resolve. A contractor with 47 subscribers who posts a detailed answer to “how much does mold remediation cost in Fayetteville Arkansas” has a realistic path to citation in Perplexity and Google AI within weeks of posting, based on the PushRank tracking data showing citation timelines of 4 days to 3 weeks from post date.
The Five-Step Video Process That Generates AI Citations
PushLeads tested this process internally during August 2026, posting YouTube videos daily throughout the month. Four of those videos were cited by AI engines. All four shared three characteristics: each answered one specific question, each had its transcript published as text on the corresponding website page, and the video was embedded on that same page. None of the four videos used paid promotion.
The five-step process is structured around those findings and requires no agency and no ad spend. Each step addresses a specific variable that determines whether a video becomes a citable source.
What is the first step in setting up a video for AI citation?
The first step is choosing one specific question per video. Not a brand overview. Not a services reel. One question a real buyer would ask, stated in the video title and in the first 30 seconds of the video. An example for a restoration contractor in Asheville would be “How much does water damage restoration cost in Asheville NC?” The city name and the cost question belong in the title, in the spoken introduction, and again in the closing summary. That repetition is what makes the transcript machine-readable for location-specific queries.
How does posting the transcript on the website increase citation probability?
AI engines index written text more consistently than they parse video metadata. When the full transcript of a video is published as readable text on the same page where the video is embedded, the page becomes a dual-format citation target. Google’s crawlers index the text, which feeds into AI Overviews. Perplexity reads the on-page text directly. The result is that one piece of content works across two citation surfaces simultaneously. This is why the AI SEO strategy of pairing video with published transcripts consistently outperforms video-only or text-only approaches in the PushRank data.
What content format does YouTube penalize or suppress?
In January 2026, YouTube removed channels with tens of millions of subscribers that were producing templated, AI-generated videos with no original face, no local context, and no substantive answers. The common characteristics of suppressed content are: slideshow-only format with no human presenter, templated scripts with no specific local detail, and short filler videos under two minutes with no depth. Using AI tools to research a topic or draft a script outline is not penalized. Producing a fully AI-generated video with no human presence and no location-specific content is the behavior that triggers suppression, based on YouTube’s stated policy actions in early 2026.
Why a Phone Camera Outperforms a $42,000 Agency Video for AI Citations
The average agency video production project costs $42,000, according to industry benchmarks cited in PushLeads’ tracking period analysis. That cost typically produces a polished brand video with professional lighting, color grading, and motion graphics. It rarely produces a video structured around a single buyer question with a spoken price range and a city-specific statement, because agency video production is optimized for brand aesthetics, not transcript density.
A contractor filming on a job site with a phone camera, speaking directly to the lens for 10 minutes about a single question, produces a denser, more citable transcript than a polished brand video that runs 90 seconds and never states a service area or a cost figure. The AI does not evaluate lighting, color balance, or production value. It evaluates whether the transcript answers the query with specificity.
What is the actual cost to produce a citation-ready video?
The hard cost is zero dollars if the contractor uses a smartphone already in their possession. The time cost is roughly 15-30 minutes per video, including filming and uploading. Posting the transcript to the website adds another 15-20 minutes if the contractor pastes the auto-generated YouTube transcript into a page manually. At that pace, a contractor could post four citation-ready videos per month and have them indexed within a week of posting, based on the 4-day-to-3-week citation timeline observed in PushRank data.
Does production quality matter at all for retention or watch time?
Stable framing, audible audio, and adequate lighting do affect whether viewers watch past the first 30 seconds. A tripod costs under $30 and eliminates the shaky footage problem. A clip-on lapel microphone costs $20-40 and eliminates wind noise and low-volume audio. Those two purchases cover the minimum threshold for acceptable production quality. Everything above that level, studio lighting, professional editing, motion graphics, adds cost without adding citation value. The AI is reading the transcript. The human viewer needs to hear and see clearly enough to stay engaged. Both needs are met with a tripod and a microphone.
How does Google’s video search behavior reinforce the YouTube strategy?
38 percent of Google search results pages include a YouTube video result, and 80 percent of those video results come from YouTube specifically. A single video posted to YouTube with a strong, question-based title can appear in traditional Google search results, in Google AI Overviews, and in Perplexity simultaneously. That is three separate discovery surfaces served by one piece of content, at a production cost that approaches zero. No other format currently achieves that cross-surface reach for a local contractor.
What Contractors Should Avoid When Building AI Citation Visibility
Several content approaches that appear productive actually generate no AI citation value and can result in channel suppression. Understanding the failure modes is as useful as understanding the success pattern, because time spent on low-citation-value content is time not spent on citation-ready content.
The most common failure mode among local contractors who attempt YouTube is posting brand videos, testimonial compilations, or equipment showcase reels. These formats have no question structure and no transcript density. They may generate some views but they will not generate AI citations because AI engines are looking for answers, not introductions.
Why do shorts fail to generate AI citations even when they get high view counts?
Shorts are under 60 seconds, which means the transcript contains fewer than approximately 150 spoken words. That is not enough text for an AI engine to extract a specific, attributable answer to a multi-part buyer question. Shorts also have a separate indexing pathway in YouTube’s architecture that is less consistently crawled by external AI engines. They are useful for growing a channel audience, but the PushRank data from 2,794 queries shows zero Shorts appearing as cited sources across the June-September 2026 tracking period.
Is there a risk that posting AI-assisted scripts will trigger YouTube suppression?
YouTube’s January 2026 enforcement actions targeted channels producing fully automated, template-generated content with no human presenter and no original perspective. Using an AI tool to research what questions homeowners ask about roofing costs, or to outline a video script, does not meet that threshold. The distinguishing factor is human presence. A contractor who appears on camera, speaks from personal field experience, and provides location-specific details is producing original content regardless of whether a script was drafted with AI assistance. The face on camera and the specific local knowledge are what differentiate citable content from suppressed content.
How does ignoring ChatGPT’s citation behavior create a gap in visibility strategy?
YouTube-only strategies address Google AI Overviews and Perplexity but leave ChatGPT coverage at zero, based on PushRank tracking data showing no YouTube citations in ChatGPT responses during the June-September 2026 period. A contractor who wants to appear in ChatGPT responses should consider building a presence on Reddit by participating in home-improvement and local community subreddits with detailed, helpful answers. That text-based presence is the format ChatGPT draws from most consistently. A complete AI citation strategy covers both video-based platforms and text-based community platforms to address all three major AI engines.
Frequently Asked Questions
How do I know if AI is currently recommending my business?
The direct method is to open ChatGPT, Perplexity, and Google AI and type the questions a real customer in your area would ask. Use variations like “who handles roof repair in [your city]” or “best water damage company near [your city] NC.” Note whether your business name appears in any response. Running a structured SEO and AI visibility audit across a broader set of queries will give a more complete picture of citation frequency across multiple platforms and question types.
How quickly can a new video start generating AI citations?
PushRank tracking data from June through September 2026 shows citation timelines of 4 days to 3 weeks from the date a video is posted. The faster end of that range occurred when the video transcript was also published as on-page text and the video was embedded on a relevant service page on the contractor’s website. Videos posted to YouTube with no corresponding website page took longer to appear in citations, particularly in Perplexity and Google AI Overviews.
Does my website need to be optimized for AI search differently than for traditional SEO?
Traditional local SEO and AI citation optimization share some foundations, including accurate business information, service-area content, and page structure. However, AI citation optimization adds specific requirements: structured question-and-answer content on service pages, published video transcripts embedded alongside the videos, and explicit price ranges or cost estimates within the page text. AI engines favor pages that directly answer buyer questions with specific numbers over pages that describe services in general terms. A combined approach that addresses both ranking signals and citation signals produces better results than either strategy alone.
What should I say in a video to make it citable for local AI searches?
Every video should include the city or neighborhood served, stated by name rather than implied. The specific service being discussed should be named in the first 60 seconds using the language a homeowner would use, not trade terminology. A price range with context for what affects the number should appear in the middle of the video. The spoken closing summary should repeat the city name and the service type. These four elements give the transcript the geographic and topical specificity that AI engines look for when matching a query to a source.
Does having more Google reviews help with AI citations?
Reviews primarily affect Google Maps ranking and Google Business Profile authority. They are a secondary signal for some AI engines, particularly those that cross-reference business reputation when generating recommendations. However, reviews are not the primary citation driver. A business with 15 reviews and a detailed YouTube video answering a specific local question will likely earn more AI citations for that question than a business with 300 reviews and no video content. Reviews and video content address different parts of the AI recommendation calculation.
How many videos do I need to post before I see AI citation results?
The PushLeads internal test during August 2026 produced 4 cited videos from a daily posting schedule over one month. Not every video will earn a citation. The ratio depends on how well each video matches an actual buyer query and how specifically the transcript addresses that query. Starting with 4-6 videos covering the highest-value questions in a trade, such as cost estimates and process explanations for the most common jobs, gives enough content surface area to begin appearing in AI responses within 30-60 days of posting.
Is Reddit actually a viable platform for a local home-service contractor?
Reddit participation earns citations primarily in ChatGPT, which does not cite YouTube in available tracking data. A contractor in Asheville who participates in community subreddits, provides detailed answers to home-repair questions with local context, and includes relevant trade knowledge builds a text-based citation profile that ChatGPT draws from. The subreddits most relevant to home-service contractors include local city subreddits, home improvement communities, and real estate communities where homeowner questions appear regularly. Participation requires written answers of at least 200-300 words to have enough text density to function as a citable source.
Can a contractor in a smaller market compete for AI citations against larger regional companies?
Yes. The Bear Restoration case from the PushRank tracking data illustrates this directly. Bear Restoration is not a large company and operates in Northwest Arkansas, not a major metro. It earned 24 citations out of 142 tested queries with no paid promotion. The citation mechanism favors specificity and relevance over company size. A small contractor in Hendersonville or Waynesville NC who posts a detailed video answering a question about roofing costs in that specific area will likely outperform a large regional company that has no local-specific video content addressing that question.
Start Showing Up in AI Answers Before Your Competitors Figure This Out
Only 1.2 percent of local businesses are currently being named by AI engines when a homeowner asks for a service recommendation. That number is not going to stay low. As more contractors recognize that AI is now handling 45 out of every 100 local business searches, the competition for citations will increase. The contractors who act now, posting structured 8-to-15-minute videos on single buyer questions, publishing transcripts as on-page text, and embedding videos on relevant service pages, will build citation authority before that window narrows.
The cost is a phone, a tripod, and a microphone. The time is 15-30 minutes per video. The citation timeline is 4 days to 3 weeks. No ad spend is required. No agency contract is required. What is required is showing up on camera, stating your city, naming your service, and quoting a price range, then repeating that process once a week.
For contractors who want to know exactly where they stand right now across ChatGPT, Perplexity, and Google AI, a structured audit is the starting point. PushLeads offers a free SEO and AI visibility audit that shows which queries are returning citations and which are not. Contractors ready to build a systematic content strategy can learn more about the full AI SEO approach or review the local SEO foundation that supports citation-ready content. Run the audit, identify the gaps, and post the first video this week.
Full Transcript
The complete spoken content of the video above, in text.
Hey, I’m Jeremy with PushLeads helping small businesses get more leads without struggle and frustration Your next customer is asking a robot the real question is is that robot? Is that AI chatbot mentioning your name? So we’re gonna dig into this.
Let’s find out so last year almost nobody was using AI to sort of find local service businesses and And it was around five or six percent in 2025 where people that were using AI to find local businesses and that moved up to 45 in 2020 and And so 45 out of 100 people are using AI to find local businesses.
So that’s not a trend That’s a stampede and this comes from the BrightLocal survey as of February 2026 Now Google still is number one.
It just has a smaller market share, right? So Google’s share of how people find a local business drop from 83 percent down to 71 percent in one you So that’s kind of a pretty big drop right there And 31 percent of people are now using ChatGPT and 23 percent are using Google’s AI mode So it’s the same customer as but a new front door And every eye is changing everything and you just got to realize that change is in the air Now being number one on Google used to be everything right something Less than half of businesses that rank well in Google maps show up when someone asks AI for a recommend date So you have two separate games happening here You have the Google maps game which is ranking number one here gets you seen by real people searching But AI assistants aren’t really looking at Google maps And then of course you have AI recommendations where one out of 80 local businesses get named and so Google’s ranked doesn’t help there Now of those local businesses ChatGPT only recommends one two of those local businesses ChatGPT is only recommending 1 .2 percent So only 1 .2 percent of those people actually come to websites, which is really Really really sad right so Google ranking is kind of invisible to AI You’re going to need a different strategy going forward and everyone knows about the click apocalypse right the clicks are drying up When Google shows an AI answer at the top most people get the answer and move on And that means that they’re getting no clicks no visits no chance to people come to the website And so it’s like 8 percent Essentially of clicking the website but if it’s an AI answering it’s like 1 percent right And that’s from the pew research center, which is looking at 68 ,000 real searches So where does the robot get all of his answers? And I was really surprised to find out it was youtube youtube is the single most quoted source In Google’s AI answers.
It’s not even close to reddit wikipedia or other sites So if you just look at this one example here 21 sources of AI answers was youtube versus eight on reddit and like six and wikipedia and three and other sites And this is from Ahrefs study in june of 2026 where they looked at 3 million searches In youtube was cited 21 percent of the time ahead of reddit wikipedia and everyone else So huge and it’s not even close across ChatGPT Perplexity and Google’s AI YouTube gets quoted 200 times more than any other video platform So look at that here that share of citations is like 20 versus vimeo and like tiktok It’s like 0 .1 and that’s from BrightEdge looking at 30 million sources And of course we didn’t take their word for it So we have our own AI tracker, which is tracking real data real clients or whatever And our tracker looked at around 2800 real Questions from our own clients tracking every answer every citation Tation every mention across ChatGPT Perplexity and Google’s AI and Here’s what k2794 real questions asked 42 42 clients tracked across all of her industries and we ended up with 813 unique question testing visibility scenario So looking at our own numbers YouTube is showing up 536 times more than anywhere else, right a youtube video was the source 536 times more to answer a question about our client industry And the breakdown was 421’s youtube citations and Google’s AI overviews brought in 421’s citations And Perplexity was 115 and that’s from our PushRank live tracker from june through september of 2026 here’s some real numbers from one of our clients bear restoration, which is a restoration company in Northwest arkansas they’re not famous.
They’re not huge out of 142 answers A real customer would ask bear got named 24 times as the answer by AI engines So one out of six question was answered by them with zero paid ads and zero SEO agency ads And the breakdown there was Perplexity was 17 answers Google AI overviews was like four And ChatGPT was like three and so the honest take here is that video is a monster to some AI engines And completely invisible to others.
So for Google’s AI and Perplexity Monster because Google drives because the video drives citations surface area and authority This is where video content earns its keep for AI search Now if you’re talking about ChatGPT and Gemini zero I mean zero citations in our live data zero doesn’t move the needle at all And so if you really want to show up for ChatGPT and Gemini or at least for ChatGPT You need to be doing Reddit right so anyone tells you that video fixes everything is selling you something the real data is split here And one thing to remember is that these robots these AI chatbots aren’t watching our videos on youtube They’re reading the transcript.
So an AI can’t sit down and watch 10 minutes of footage But it reads the transcript every word you said written down So what you say out loud is when an AI is going to quote So you need to speak in your videos, right? If you don’t say anything in the video then the AI doesn’t even know you said anything And so you’ll never get cited, right? So it’s important in the video to say your town your city your neighbor your service Do you what do you do? You do roof repair HVAC install kitchen remodeling and then say the price It’s important to say the prices or at least a range in prices because AI is really loves specifics And then another thing to remember is that you don’t have to be famous Right it doesn’t even matter if you’re TikTok famous or not Actually 41 % of AI quoted youtube videos have under a thousand views And that’s a that’s a huge number that’s about half the videos So you don’t have to have a channel with like five or ten or a hundred thousand subscribers You just get to put out really good quality content that Google and ais will cite And so really the truth is it’s not really about the channel size.
It’s about the quality of the content It’s really always about the quality of the content But the best answer wins not the biggest channel and you don’t have to be rich, right? An agency video can cost 42 thousand dollars per average project which is polished expensive And often says nothing as far as answering questions that and somebody would actually search using an AI or Google But you could do and your phone beats it every time because it’s zero Dollars and if you actually answer the question and give value and retention And talk for the entire five or ten minutes The robot doesn’t care about your lighting.
It cares whether or not you’ve answered the question And so if you’re answering the questions, you’re going to win over these agency videos that are like A hundred or a thousand times more expensive.
So are we actually in the game? Or do we practice what we preach? Yes, we do we’ve been posting youtube videos every day Throughout all of august and four of our own videos got quoted by AI And every single one of these videos had three things in common nothing else Each one answered one specific question Not a brand overview not a demo reel One specific question that your customer actually asked secondly the transcript was posted as text on the website Why because the robots read the text right the transcript made the text indexable and quotable And then thirdly we embedded the video on the matching page of the site The video lived on the same page on the site that answered the same question And so and so time from being posted to being quoted by AI’s was four days to three weeks That’s it with no ad spend no agency just in the right place Answering the question And so here’s the recipe five steps no agency no ad spend in the right place One question per video how much does water damage cost in Asheville? Whatever your question is say your city Asheville North Carolina.
That’s a quotable line Add the transcript in the chapters.
This is the part that AI reads Make the videos eight to 15 minutes because 94 percent of what AI quotes goes to these long videos not shorts Shorts will bring people to your channel, but they’re not going to actually earn AI citations And of course then put the video right on your page with the words the transcript below that super important you do that And what gets you buried is low quality slide show only Like AI slop basically so in january 2026 youtube shut down channels that had tens of millions of subscribers And these channels were producing empty temp empty templated content And this is not theoretical youtube really did shut these down And so it’s fine to use AI tools to help you script an AI to research answers But what gets you buried is using templated videos with no real face Math produced content with no local context and shorts and filler with no substance So really so really the bottom line is your face on a job site or on location giving the real answer That’s the whole trick And of course we know that Google still shows youtube videos as well right 38 percent of results actually show youtube results In the search right and so nearly four out of ten searches surface of video result And then 80 percent of video searches come from youtube youtube dominates in video search by far And so the truth is that one video can work in the old Google and in the new AI realm at the same time And this is the best and bill in marketing right now to answer One question in one video and make that video Eight to fifteen minutes long because it’s going to be citable versus like a short right So really there’s a huge opportunity here.
The door is wide open Only 1 .2 percent of local businesses are getting recommended by AI today And almost nobody across all these media industries are actually doing youtube This means the opportunity is huge right now So when someone asks AI who’s the best SEO company in town or who’s the best water damage specialist in Asheville or whatever your city is Let’s do the thing make the videos and make sure that AI is answering you and you get recommended And you get that visibility.
All right, i’m jeremy ashburn with PushLeads helping small businesses get more leads without struggle and frustration Talk to you soon
Keep Reading
Want more customers from Google & AI search?
Get a free SEO audit of your site — see exactly what to fix first.