35 answers from Jolly Search on what gets a brand named by ChatGPT, Google AI Overviews, Google AI Mode, Gemini, Perplexity, Claude and Grok, what does not, how long it takes, and what it costs.
Published September 21, 2026 · Last updated October 6, 2026
AI visibility is whether AI engines name your brand in the answer when buyers ask a question in your category, on ChatGPT, Google AI Overviews, Google AI Mode, Gemini, Perplexity, Claude and Grok. Brands that get recommended target buying prompts rather than "what is" prompts, have several independent sources that agree about them (Reddit, YouTube, LinkedIn, earned media, their own website), and run a website that holds up when an AI engine checks who they are [S1][S2][S13].
4 questions
AI visibility is how often, and how favorably, AI engines name a brand when buyers ask questions in its category. The engines that matter most today are ChatGPT, Google AI Overviews, Google AI Mode, Gemini, Perplexity, Claude and Grok.
The goal is not traffic. It is being one of the handful of names in the answer. In Jolly Search campaigns, AI answers shortlist four or five brands at most, so a brand that is not named is never evaluated at all [S13][S2]. As Muskaan Rana, who runs AI visibility production at Jolly Search, puts it: a brand that is not recommended for its category query is invisible to its buyers [S13].
Cited means an AI engine used your content as a source. Recommended means the AI engine named your brand in the answer itself. A citation is a footnote, a recommendation is a shortlist [S2].
| Cited | Recommended | |
|---|---|---|
| Where the brand appears | Sources panel, link chip, or footnote list | The answer itself, as a named option |
| What the AI engine is doing | Borrowing the content as evidence | Putting its credibility behind the brand |
| Typical prompt type | Informational: "What is [service]?" | Buying: "Which [service] for [buyer]?" |
| What the buyer does next | Rarely clicks the footnote | Opens a new tab and searches the brand |
Most brands sit on one of three rungs for any given prompt: absent (the AI engine does not know the brand exists for that question), cited but not recommended (it trusts the data but not the brand), or recommended [S2]. Rung two is the trap, because it looks like progress in a report without putting the brand in front of a buyer.
Rung two can be closed quickly. Google AI Overviews cited an online oral-care brand's website without recommending it. Within two weeks of one buying-guide article answering the Target Prompt, the same engine recommended the brand as Best Overall [S23].
The engines your buyers actually use, and they do not all behave the same way. Jolly Search tracks ChatGPT, Google AI Overviews, Google AI Mode, Gemini, Perplexity, Claude and Grok on every campaign, and adds surfaces such as Microsoft Copilot or Reddit's own AI answers where they matter [S1][S15].
Engines can agree. For an IoT engineering firm, Google AI Overviews, Google AI Mode, ChatGPT, Perplexity, Gemini, Grok and Reddit Answers all named the firm first for its Target Prompt on the same day [S28].
Results are reported engine by engine rather than as a single "AI score."
Usually one of five reasons, and more than one at a time is common.
3 questions
SEO works to rank a page, AI visibility works to get a brand named inside an AI answer. The two still overlap, and SEO fundamentals are now the entry ticket rather than the whole game. A year ago Jolly Search described generative engine optimization (GEO) as more than 90% overlapping with SEO best practices, and that overlap has shrunk [S13].
GEO, AEO (answer engine optimization), AI SEO and AI search optimization all describe the same work. At Jolly Search, GEO always means generative engine optimization, never geographic.
Yes. Technical SEO, onsite content and earned mentions on trusted publications are the foundation AI visibility is built on. AI engines verify a brand by crawling its website, so a site that fails the basics stalls the campaign before it starts [S13].
What has changed is that SEO alone is no longer enough. Getting recommended also takes off-site sources AI engines lean on for corroboration: Reddit threads, YouTube videos, LinkedIn posts and earned media [S7]. See Branded Backlinks for the earned-media side.
No. One mistake Jolly Search sees is brands dropping SEO entirely to chase social platforms and AI engines. The Reddit citation drop on ChatGPT in August 2026 showed why no single platform can carry a brand [S3].
Keep investing in the website, digital PR and off-page authority, and add AI visibility on top. Looking into 2027, the number of platforms to optimize for is growing rather than shrinking: Google, multiple AI engines, and social platforms at the same time [S13].
6 questions
AI engines follow three steps: discovery, then corroboration, then the default answer. That pattern has been consistent across Jolly Search campaigns [S7].
The crawl order when an AI engine validates a brand is also consistent: first the asset that answered the question, then the About page, then the page that proves the claim (for a retailer, the shipping page). Citing is cheap for a model, recommending is a reputational bet, so it checks first [S2].
A Target Prompt is the business-critical question a buyer types into an AI engine right before spending money. For example: "Which [specialty] equipment retailers in the US ship nationwide?" Every category has a handful of them, and whoever the engines recommend in those answers owns the demand behind them [S8].
A prompt is not a keyword. A keyword is what someone types into Google to find a page; a Target Prompt is a full question that asks an AI engine to choose between brands. Owning one Target Prompt rarely stops at one prompt, because the same work tends to win the natural variations of that question, which is why the Jolly Search service is called Own Your Category [S13].
A specialty concrete-tool retailer shows how one prompt grows into a category. After its first Target Prompt win, the campaign expanded to more than a dozen buying prompts, and for a financing prompt Google AI Overviews, ChatGPT, Perplexity, Grok and Claude all put the retailer first, ahead of Home Depot and White Cap [S26][S1].
Because the buyer did not ask for a brand. A "what is [service]?" prompt produces a lesson: a definition, examples, how it works. The best outcome for that prompt is becoming a cited source [S2].
One Jolly Search client wanted to target "What is debt collection?" The campaign moved to a buying version, roughly "What is the best debt collection agency for marketing agencies?", and the client is recommended for that prompt today [S13]. In a published Jolly Search case study, a commercial collection agency is ranked first by Google AI Overviews and Google AI Mode for the best commercial collection agency in the US [S21]. Buying prompts usually start with "which," "best" or "vs," or they name a specific buyer.
Start with the ideal customer, not the keyword list. Define the buyer persona and their customer journey first, then map the bottom-of-funnel questions that buyer asks when comparing options [S13].
Starting from keywords is fine. Jolly Search derives Target Prompts from the keywords a brand already cares about, then validates them against how buyers phrase questions to AI engines.
Yes. Content carrying genuinely new information, especially first-party data, is what earns preference. AI engines can already summarize a thousand existing sources faster than any writer, so a generic informational article adds little. What they cannot produce is a company's own data and first-hand experience [S13].
An article built on data only one company holds makes that company a credible source for informational questions and strengthens its case to be recommended for commercial ones [S13].
AI-written text carries a risk worth managing, which is why Jolly Search AI visibility assets are written by people on the team. Since August 2026, text from newer Claude models carries an invisible watermark, and Google, Meta, Microsoft and OpenAI have signed the same EU code of practice [S9].
No public detector flags content today. The Reddit threads, LinkedIn posts and YouTube videos that get brands recommended are exactly the surfaces a platform would label first if one ships [S9]. Use AI for drafts, and keep opinions, first-hand data, client stories and voice human [S9][S13].
8 questions
Yes, though Reddit should never be a brand's only channel. In mid-August 2026, Reddit's share of ChatGPT citations fell from about 3.8% to 0.5% in a week [S3][S10]. Google barely moved: Reddit's share in Google AI Overviews went from 2.4% to 2.1% [S3].
ChatGPT also did not stop reading Reddit. Reddit stayed at roughly 25% to 34% of the pages ChatGPT consulted while citations collapsed [S11]. In a free test after the drop, one Reddit thread took a law firm from absent to named in Google's AI Overview in about three weeks [S3].
Reddit still works fast on Google. In 2026 Jolly Search case studies, Reddit discussions were cited by Google AI Overviews, Google AI Mode and Reddit Answers within 1 to 3 days of going live [S27], and Reddit discussion put a kitchen and bath remodeler in 14 of 14 readable AI answers for its local prompts [S25].
Reddit is the discovery signal. Getting recommended on ChatGPT and Gemini takes corroboration from the brand's website, YouTube, LinkedIn and third-party mentions [S3][S13]. See Elite Reddit Branded Posts.
Posts get removed when they read like promotion from accounts that have not earned the right to promote. Across 2,670 comments in 411 subreddits, the Jolly Search deletion rate fell from 74% in January 2026 to 2.4% in May 2026, below Reddit's sitewide removal rate of about 7% [S6].
Two things did it: leading with genuinely useful contributions and integrating a brand only where it earns its place [S6], and choosing communities for long-term survival rather than size [S17]. For AI visibility, a thread that stays live and indexed is worth far more than one in a bigger subreddit that gets removed [S15].
Yes, especially on Google and Perplexity. A 9-minute video Jolly Search built for AI engines rather than human viewers went live on a client's channel on August 5, 2026. By August 7 it was in the source panels of Google AI Overviews and Google AI Mode, and Perplexity cited it by URL [S4].
See YouTube for AI Visibility.
Yes. LinkedIn posts and articles get cited, and third-party social proof helps AI engines connect a brand to its category. For an HR software marketplace client, a LinkedIn post was cited in Google AI Overviews and Perplexity, while the same client's Reddit thread and onsite blog earned citations on other engines [S7]. On another prompt, a LinkedIn post that was three days old was cited by Google AI Overviews, which listed the client first, as best for buyer support, above G2 and Capterra [S20][S18].
It works in narrow niches too. For a managed bookkeeping service, Google AI Overviews and Google AI Mode cited the client's LinkedIn post alongside its onsite article for a payment reconciliation prompt [S33].
Tracked 11 days apart, that client stayed the number one recommendation for its money prompt, with the AI engine citing the LinkedIn post the second time [S2]. Social and search are converging for a reason: when other people describe a company as the expert, AI engines associate the brand with that category over time [S13]. See LinkedIn for AI Visibility.
Not as a core channel. In Jolly Search prompt tracking, Quora is cited far less often than it used to be, so AI visibility campaigns are no longer built around it. Reddit, YouTube, LinkedIn and a brand's own website show up far more consistently in AI answers [S15].
Because a brand's website is the one source that cannot be turned off. Third-party citations can drop at any time, as Reddit's did on ChatGPT within a week in August 2026 [S3].
Onsite results are measurable. For a land-investing client, Google Search Console's generative AI report showed AI impressions up 88% in two months and homepage AI impressions up 105%. The one article written specifically for AI visibility became the third-highest page on the site by AI impressions, and in regular search it grew 483% in 28 days, outpacing 24 articles written for traditional SEO [S3][S22][S19].
Onsite articles also get cited fast. An AI video-to-music tool's article was cited by Google AI Mode, Gemini, Perplexity and Claude nine days after publishing [S32], and a baby-bottle brand's own store page was cited when Google AI Overviews began recommending it [S24]. See onsite content.
Yes. A Jolly Search client's blog post answering its buyers' most valuable question showed "URL is not on Google" in Google Search Console, yet the same page was cited by ChatGPT, Grok, Perplexity and Google AI Overviews [S5].
Zero Google clicks does not mean a page is doing nothing, because AI crawlers read pages without clicking them [S5].
Anything that stops AI crawlers reading a website, or makes the site look unreliable once they do. These are the issues Jolly Search finds most often in technical AI visibility audits [S15][S13]:
7 questions
By asking the Target Prompts on every engine, on a schedule, and recording whether the brand is absent, cited or recommended. The proof is the Day 0 answer compared with today's answer, engine by engine [S3].
Jolly Search has published AI visibility results in SaaS, finance, manufacturing, e-commerce, professional services, education, legal, construction and healthcare. Every result below is screenshot-verified on the live AI engine, and each client is anonymized [S35].
| Industry | Documented result | Source |
|---|---|---|
| SaaS | HR software marketplace listed first by Google AI Overviews, above G2 and Capterra | [S20] |
| SaaS | Land data platform grew Generative AI impressions 88% in Google Search Console | [S22] |
| Finance | Commercial collection agency ranked first by Google AI Overviews and AI Mode for its category | [S21] |
| Finance | Bookkeeping service recommended first by Reddit Answers, with its article and LinkedIn post cited by Google AI | [S33] |
| Manufacturing | Contract manufacturer recommended first for precision CNC parts, above PCBWay | [S29] |
| Manufacturing | Components distributor cited within 1 to 3 days, in English and German | [S27] |
| E-commerce | Specialty tool retailer best overall in 5 of 5 AI engines for a financing prompt | [S26] |
| E-commerce | Oral-care brand named Best Overall by Google AI Overviews on day 14 | [S23] |
| Professional services | IoT engineering firm named first by seven AI surfaces on the same day | [S28] |
| Education | Healthcare training provider recommended first by Google AI Overviews, above Penn Foster | [S30] |
| Legal | UK law firm recommended first by Reddit Answers and named a top-tier firm by Google AI Overviews | [S34] |
| Construction | Kitchen and bath remodeler named in 14 of 14 AI answers for its local prompts | [S25] |
| Healthcare | Telehealth clinic recommended first by Grok and Reddit Answers | [S31] |
The pattern holds across categories: buyers ask an AI engine for a shortlist before they visit a website, and the brands named in that answer are the ones evaluated [S2].
Yes. AI engines answer the exact question asked, so a specialist with a specific, current source can be named ahead of much larger brands. These are documented Jolly Search results:
Head-on competition with a dominant brand still takes more corroborating sources, so these wins come from choosing prompts where the specialist is the better answer [S7].
Yes. Homeowners and local buyers now ask AI engines who to hire, and local prompts can be won outright. A family-owned kitchen and bath remodeler appeared in 14 of 14 readable AI answers across 4 local prompts and 8 AI engines, ranked first in 9 of them, including ChatGPT's best overall pick [S25].
Local prompts have small volume, but each one is a buyer ready to hire. ChatGPT, Google AI Overviews and Reddit Answers cited the remodeler's Reddit discussion directly [S25].
Yes. AI engines answer in the buyer's language, and the same approach that wins an English prompt can win the local-language version. For an electronic components distributor selling into Europe, a German-language Reddit discussion was cited by Google AI Overviews three days after posting [S27].
The same client was named by ChatGPT, Gemini, Google AI Overviews, Google AI Mode and Reddit Answers for sourcing prompts in English [S27].
No, not since mid-August 2026. The citation list is no longer a complete record of what an engine read. Trackerly found Reddit stayed at roughly one in four pages ChatGPT consulted while Reddit citations fell to about 0.5% [S11][S3].
Jolly Search saw the same pattern on Google: a thread appeared in the source panel one day and dropped out the next, while the recommendation stayed. That is why a citation is no longer treated as proof of results, and Day 0 versus today is [S3].
Because AI crawlers read without clicking, and many AI apps strip the referrer. Visits from AI apps often land in analytics as direct traffic [S5].
Bots also make up the majority of page requests now: 60.5% of HTML requests in the week of August 11 to 17, 2026 [S5]. Watch branded search impressions and AI crawler activity alongside referral traffic to see the full picture [S5].
2 questions
Days to weeks for the first movement, which is much faster than classic SEO. Timelines vary by category and competition. These are results Jolly Search has documented:
| Documented result | Time | Source |
|---|---|---|
| Reddit Answers recommends a healthcare training provider first for its certification prompt | 1 day | [S30] |
| Reddit discussions cited by Google AI Overviews, Google AI Mode and Reddit Answers, in English and German | 1 to 3 days | [S27] |
| YouTube video appears in Google AI Overviews and Google AI Mode source panels, and Perplexity cites it by URL | About 48 hours | [S4] |
| Onsite blog post answering a Target Prompt is cited by ChatGPT, with tagged referral traffic | Within 72 hours | [S8] |
| Reddit thread ranks #1 on Google for the Target Prompt itself | 3 days | [S8] |
| A three-day-old LinkedIn post is cited by Google AI Overviews, which lists the client first, above G2 and Capterra | 3 days | [S20] |
| Baby-bottle brand goes from absent to a recommended product in Google AI Overviews | 7 days | [S24] |
| Onsite article cited by Google AI Mode, Gemini, Perplexity and Claude | 9 days | [S32] |
| Oral-care brand goes from cited to "Best Overall" in Google AI Overviews | 14 days | [S23] |
| Client goes from absent to recommended on all seven AI surfaces Jolly Search tracks | 19 days | [S1] |
| One Reddit thread in a free test: law firm goes from absent to named in Google's AI Overview | About 3 weeks | [S3] |
Competitive categories take longer, because displacing a well-known brand takes more corroborating sources [S7].
Only with continued investment. AI answers are living documents. In a six-week test, a client's visibility rose, disappeared when the campaign paused, and returned when it resumed [S5].
Competitors contest answers too. In one category, a larger rival follows a Jolly Search client's Reddit threads and joins them to promote itself [S16]. Assets are maintained rather than published and left, and Own Your Category includes defending a prompt after a brand owns it.
5 questions
Jolly Search AI visibility campaigns run from $697 to $14,997 per month, depending on how many prompts and channels a brand needs. Every service is a monthly subscription, and all prices below are per month.
| Service | Monthly price | What it covers |
|---|---|---|
| Own Your Category, Authority | Custom | Category-wide prompt coverage with competitive defense |
| Own Your Category, 3 prompts | $14,997 | Three Target Prompts, adding LinkedIn, YouTube, and Google AI Overviews and AI Mode tracking |
| Own Your Category, 1 prompt | $4,997 | One Target Prompt: Reddit threads, onsite content, citation tracking, monthly ownership reporting |
| YouTube for AI Visibility | $6,997 / $3,797 / $1,997 | 20, 10 or 5 AI-optimized videos per month, scripts included, no setup fee |
| Elite Reddit Branded Posts | $4,497 / $3,497 / $2,497 | 8, 6 or 4 actively maintained Reddit assets per month |
| LinkedIn for AI Visibility | $2,377 / $1,277 / $697 | 20, 10 or 5 LinkedIn assets per month on your company page |
Onsite content is priced per article on the content page, and earned Branded Backlinks are priced per link.
Own Your Category is the Jolly Search AI visibility service, measured by one thing: whether AI engines recommend a brand when its Target Prompt is asked. A campaign combines Reddit threads, onsite content, YouTube, LinkedIn and citation tracking into one program built to win and hold the answer [S1][S4].
Pricing runs $4,997 per month for one Target Prompt, $14,997 per month for three prompts, and custom pricing for category-wide Authority coverage. Full detail is on the Own Your Category page.
Yes. Every Jolly Search channel is sold on its own as a monthly subscription. YouTube for AI Visibility runs $6,997, $3,797 or $1,997 per month, Elite Reddit Branded Posts $4,497, $3,497 or $2,497 per month, and LinkedIn for AI Visibility $2,377, $1,277 or $697 per month.
One channel rarely gets a brand past cited when it is competing with established names, which is when a multi-channel campaign pays off [S7].
No. Nobody can honestly promise what an AI engine will say. Jolly Search commits to the work set out in the agreement, and reports the outcome that matters, prompt ownership, engine by engine with Day 0 versus present-day evidence [S2][S3].
When an agency does promise AI recommendations, ask how it defines recommended and how it proves the result.
Run your money prompt through ChatGPT and Google AI Mode and note whether your brand is absent, cited or recommended. If the prompt starts with "what is," find the buying version of it first [S2].
Jolly Search will also run the check for you: get a free AI Visibility Report, or book a strategy call and the team will test your prompts live and tell you which rung your brand is on.
References
Bring us your money prompt. We will run it across the AI engines live and show you exactly where your brand stands.
Book a Strategy Call →