# How to get your business recommended by ChatGPT and AI search in 2026

URL: https://yarigai.ai/guides/get-recommended-by-chatgpt
Author: Deniz Sancar, founder of Yarigai (https://denizsancar.com)
Publisher: Yarigai, a done-for-you AI agency in London, UK (https://yarigai.ai)
Published: 7 November 2026. Updated: 7 November 2026.

## Short answer
To get a business recommended by ChatGPT and AI search in 2026, let each engine's search crawler in, answer the buyer's question with facts and prices in the first lines of each page, and earn mentions on other sites. In the GEO study, adding statistics, quotations or cited sources raised a page's visibility in AI answers by about 27% to 41%. Ahrefs found branded web mentions track Google AI Overview presence three times more closely than backlinks. llms.txt shows no measured effect.

## At a glance
| Finding | Figure | Source |
|---|---|---|
| Visibility lift from adding quotations, statistics or sources | About 27% to 41% over an unchanged page | GEO paper, KDD 2024, 10,000 queries |
| Keyword stuffing | Below the unchanged page; 10% worse on Perplexity | GEO paper |
| Branded web mentions and AI Overview presence | 0.664 correlation, against 0.218 for backlinks | Ahrefs, 75,000 brands, May 2025 |
| Cited page types for "best X" prompts | 43.8% were "best" lists; agencies' own service pages appeared in 30.4% of agency prompts | Ahrefs, 750 prompts, December 2025 |
| Presence on several review platforms | 4.6 to 6.3 average ChatGPT citations, against 1.8 without | SE Ranking, 129,000 domains, November 2025 |
| llms.txt files that got no requests in a month | 97% | Ahrefs, 137,210 domains, June 2026 |
| Claude citations matching Brave Search results | 79.2% across about 400 queries | Profound, June 2026, reported by xponent21 |
| Same brand list on a repeat run of a prompt | Under 1% of the time on ChatGPT and Google AI Overviews | SparkToro and Gumshoe, reported by Search Engine Journal, January 2026 |

Most of these studies come from companies that sell search or AI-visibility tools, and they measure correlation, not cause.

## Which index each AI engine searches
Each AI engine answers from its own index or a partner's, so a page has to be open to several sets of crawlers. OpenAI and Anthropic run separate crawlers for search and for model training, so a site can stay in search answers while it opts out of training.

| Engine | Where answers come from | Crawlers to allow | What the engine's docs say |
|---|---|---|---|
| ChatGPT search | OpenAI's own crawler plus outside search providers | OAI-SearchBot, and ChatGPT-User for live visits | Sites opted out of OAI-SearchBot "will not be shown in ChatGPT search answers"; robots.txt changes take about 24 hours |
| Perplexity | Perplexity's own index | PerplexityBot and Perplexity-User | Perplexity-User "generally ignores robots.txt rules"; changes take up to 24 hours |
| Claude | Brave Search results, plus Anthropic's own crawlers | Claude-SearchBot and Claude-User, and Googlebot for Brave | Blocking Claude-SearchBot "may reduce your site's visibility and accuracy in user search results" |
| Google AI Overviews and AI Mode | Google's Search index | Googlebot | No new machine-readable files, markup or Markdown needed |
| Microsoft Copilot | Bing's index | Bingbot | Bing Webmaster Tools reports Copilot citations |

Three details matter for a small business site:
- OpenAI's crawler page names OAI-SearchBot as its search crawler but does not say which outside providers ChatGPT search uses or how much each contributes. Getting into Bing's index costs nothing and also covers Copilot, so it is the sensible first step.
- Anthropic added Brave Search as a web search subprocessor on 19 March 2025, xponent21 reports. Brave's crawler won't crawl a page that Googlebot is blocked from, so a Googlebot block also cuts a site out of Claude's main source.
- Perplexity warns that Cloudflare's Bot Fight Mode can still challenge its bots even when a custom rule allows them. A site on Cloudflare should check its bot settings and its logs for blocked crawler visits.

## What the studies found
The strongest evidence says pages get quoted when they carry statistics, quotations and named sources, and brands that other sites mention get recommended more often. Page format matters less than either.

### The GEO paper
Aggarwal and colleagues tested nine ways of rewriting web pages against GEO-bench, a set of 10,000 queries, and published the results at KDD 2024. Against a baseline score of 19.3 on their position-adjusted word count, adding quotations scored 27.2 (about 41% higher), adding statistics 25.2 (about 31%) and citing sources 24.6 (about 27%). Keyword stuffing scored 17.7, below doing nothing. Citing sources helped low-ranked pages most. It lifted visibility by 115.1% for pages ranked fifth, while the top-ranked page fell by 30.3%. The paper was first posted in November 2023, so it tested engines from that year.

### Ahrefs on brand mentions and cited pages
Across 75,000 brands, Ahrefs found that branded web mentions correlated 0.664 with presence in Google AI Overviews, ahead of branded anchors at 0.527, branded search volume at 0.392 and backlinks at 0.218 (May 2025). In a December 2025 study of ChatGPT answers to 750 prompts, with 26,283 source URLs, "best X" blog lists made up 43.8% of cited page types. For agency prompts, an agency's own service landing page was cited in 30.4% of prompts, ahead of the agency's own blog lists at 17.2% and its homepage at 14%. Of the cited lists, 79.1% had been updated in 2025.

### SE Ranking on ChatGPT citations
SE Ranking modelled 216,524 pages on 129,000 domains across 20 niches (November 2025). Domains with up to 2,500 referring domains averaged 1.6 to 1.8 ChatGPT citations, against 8.4 for those with over 350,000. Pages with expert quotes averaged 4.1 citations against 2.4 without, pages with 19 or more statistics 5.4 against 2.8, and pages updated within three months 6.0 against 3.6.

### The llms.txt log study
Ahrefs used its Bot Analytics tool to check every May 2026 request for llms.txt across 137,210 domains. 28% published an llms.txt file, 97% of those files got no requests at all, and retrieval bots such as OAI-SearchBot made about 1.1% of the requests. The author's verdict was "the cons outweigh the pros right now."

## What doesn't work
Tactics aimed at the engine rather than the buyer show no lift in the studies above, and Google's own guide warns against several of them.

- Keyword stuffing. It scored below an unchanged page in the GEO paper and did 10% worse on Perplexity.
- llms.txt as a ranking lever. Google's guide says Google Search ignores it, and SE Ranking's model predicted citations better with llms.txt left out.
- Schema markup as a citation lever. Google says structured data "isn't required for generative AI search". In SE Ranking's data, pages with FAQ schema averaged 3.6 citations against 4.2 without.
- A page for every wording of a query. Google says doing this to manipulate AI responses breaks its scaled content abuse policy.
- Bought or fake mentions. Google says "seeking inauthentic 'mentions' across the web isn't as helpful as it might seem."
- Cutting pages into tiny pieces. Google says there is no requirement to break content up for AI.

## A step-by-step plan for a small business
Work through these in order. The first four are on your own site and take days. The fifth takes months.

1. Let the search crawlers in. Allow OAI-SearchBot, ChatGPT-User, PerplexityBot, Perplexity-User, Claude-SearchBot, Claude-User, Googlebot and Bingbot in robots.txt. Training crawlers such as GPTBot and ClaudeBot are a separate choice. Expect up to 24 hours for a change to reach OpenAI and Perplexity.
2. Give each service its own page that answers the buyer's question in the first lines. State who you are, where you work, who you serve, what it costs or the price range, and the date. When ChatGPT cites an agency's own site, it cites a service landing page more often than the homepage or a blog list.
3. Publish your facts and prices as plain text. Include your company number, address, services, delivery times, terms and prices or ranges. An engine can only quote a number that is on the page.
4. Add statistics, quotations and named sources to each page. These were the three strongest changes in the GEO paper.
5. Earn mentions and reviews on other sites. Ask customers for reviews on the platforms buyers in your sector read, and send your facts to the authors of honest "best X" comparisons. Mentions correlated with AI Overview presence three times more strongly than backlinks in the Ahrefs study.
6. Get into Bing. Verify the site in Bing Webmaster Tools, submit the sitemap and switch on IndexNow so Bing hears about new pages as you publish them.
7. Get into Google. Verify the site in Google Search Console, check that each service page is indexed, and use its Generative AI performance report.
8. Keep pages current. Pages updated within three months averaged 6.0 ChatGPT citations against 3.6 in SE Ranking's data, so show a visible updated date and change it only when the content changes.
9. Measure your share of answers every month, as the next section sets out.

## How to measure share of answers
Count how often the business appears across many runs of many prompts, not where it ranks in a single answer. SparkToro and Gumshoe ran 12 prompts 60 to 100 times each on ChatGPT, Claude and Google's AI results. ChatGPT and Google's AI Overviews each returned the same brand list less than 1% of the time, and the same list in the same order came back less than 0.1% of the time, Search Engine Journal reports. Claude repeated its list slightly more often.

The set of brands considered is steadier than the list. When 142 people wrote their own headphone prompts, their wording had a similarity score of only 0.081, yet the top four brands appeared in 55% to 77% of 994 answers.

A simple monthly method:
1. Write 20 to 40 prompts in your buyers' words, such as "best accountant for a small charity in Leeds" or "who builds booking systems for clinics".
2. Run each prompt at least 10 times on each engine, logged in and logged out.
3. Score the share of answers that name you and the share that link to your site. A prompt run 10 times on 4 engines gives 40 answers, so a business named in 12 of them has a 30% share for that prompt.
4. Check the first-party reports. Bing Webmaster Tools' AI Performance report, in public preview since February 2026, shows total citations, cited pages and grounding queries across Copilot and Bing's AI summaries, but no clicks or traffic. Google recommends the Generative AI performance report in Search Console.
5. Read your server logs for visits from OAI-SearchBot, PerplexityBot, Claude-SearchBot and Claude-User. A visit shows an engine retrieved the page before it cited anything.

Treat third-party trackers as samples. Google says "No third-party tool has access to our internal ranking or AI systems."

## How Yarigai builds pages that rank in Google and ChatGPT
Yarigai builds pages that rank in Google and ChatGPT as part of its [AI marketing systems](/services/ai-marketing-systems), alongside [websites, lead capture and booking flows](/services/websites-and-apps). It builds inside the client's own accounts and tools, and the client owns the code, the data and the accounts from day one.

A project starts with a free 30-minute scope call on Google Meet, and the client keeps the notes whether they hire Yarigai or not. A blueprint of about a week, credited to the build, sets out a written plan, a clickable prototype, the monthly running cost, a fixed price and the payback maths. Yarigai's guarantee reads "If the build isn't profitable for you, we won't build it." Nobody can guarantee a citation, so judge any supplier, Yarigai included, on the pages it ships and the share of answers you measure yourself. For what a site itself costs, see the [small business website cost guide](/guides/website-cost-uk).

## Common questions
### Does llms.txt help ChatGPT find my site?
None of the studies in this guide found that it does. Ahrefs found that 97% of llms.txt files on 137,210 domains got no requests in May 2026, and Google says the file neither helps nor harms a site in Google Search. Crawler access and clear pages matter far more.

### Should I block GPTBot to stop my content training AI models?
You can, and keep OAI-SearchBot allowed. OpenAI says GPTBot is its training crawler and that sites opted out of OAI-SearchBot, its search crawler, will not be shown in ChatGPT search answers. Anthropic splits ClaudeBot for training from Claude-SearchBot and Claude-User for search in the same way.

### How long does it take for ChatGPT to cite a new page?
No engine publishes a figure. OpenAI says robots.txt changes take about 24 hours to reach its search systems, and Perplexity says up to 24 hours. Submitting new pages through Bing Webmaster Tools and IndexNow tells Bing about them on the day you publish.

### Does schema markup help with AI search?
Not much, on current evidence. Google says structured data isn't required for its AI features and there is no special markup to add. SE Ranking found pages with FAQ schema averaged 3.6 ChatGPT citations against 4.2 without, a correlation that probably reflects page type.

### Why does ChatGPT recommend my competitor and not my business?
The studies point off-site first. "Best X" lists make up 43.8% of cited page types for those prompts in Ahrefs' study, and branded mentions correlate with AI Overview presence far more than backlinks do. Check which lists and review sites name your competitor, and make sure your own service pages state your facts and prices.

### Do online reviews affect AI recommendations?
They correlate with citations. In SE Ranking's study, domains present on several review platforms, such as Trustpilot, G2 and Capterra, averaged 4.6 to 6.3 ChatGPT citations, against 1.8 for those on none. The study shows correlation, not cause.

## Sources
1. [GEO: Generative Engine Optimization](https://arxiv.org/abs/2311.09735), Aggarwal, Murahari, Rajpurohit, Kalyan, Narasimhan and Deshpande, arXiv, accepted to KDD 2024, checked 7 November 2026.
2. [Overview of OpenAI Crawlers](https://developers.openai.com/api/docs/bots), OpenAI, checked 7 November 2026.
3. [Does Anthropic crawl data from the web, and how can site owners block the crawler?](https://support.claude.com/en/articles/8896518), Anthropic, checked 7 November 2026.
4. [Perplexity Crawlers](https://docs.perplexity.ai/guides/bots), Perplexity, checked 7 November 2026.
5. [Anthropic Lists Two Web-Search Subprocessors Behind Claude, and the GEO Playbook Missed the Second](https://xponent21.com/insights/claude-web-search-brave-turbopuffer/), xponent21, published 2 July 2026, updated 16 July 2026.
6. [Google's Guide to Optimizing for Generative AI Features on Google Search](https://developers.google.com/search/docs/fundamentals/ai-optimization-guide), Google Search Central, checked 7 November 2026.
7. [An Analysis of AI Overview Brand Visibility Factors (75K Brands Studied)](https://ahrefs.com/blog/ai-overview-brand-correlation/), Ahrefs, published 26 May 2025.
8. [Do Self-Promotional "Best" Lists Boost ChatGPT Visibility? Study of 26,283 Source URLs](https://ahrefs.com/blog/best-lists-research/), Ahrefs, Glen Allsopp, published 4 December 2025.
9. [How to Optimize for ChatGPT: Skip LLMs.txt, Earn Trust on Quora and Reddit](https://seranking.com/blog/chatgpt-citation-factors/), SE Ranking, published 24 November 2025.
10. [We Analyzed 137K Sites: 97% of llms.txt Files Never Get Read](https://ahrefs.com/blog/llmstxt-study/), Ahrefs, published 15 June 2026.
11. [AI Recommendations Change With Nearly Every Query: Sparktoro](https://www.searchenginejournal.com/ai-recommendations-change-with-nearly-every-query-sparktoro/566242/), Search Engine Journal, published 30 January 2026.
12. [Bing Webmaster Tools Launches AI Performance Reporting](https://www.koozai.com/blog/search-marketing/bing-webmaster-tools-launches-ai-performance-reporting/), Koozai, published 13 February 2026.