Prices checked monthlySingapore event vendors, checked.

Corporate events / AI visibility

How to get recommended by ChatGPT: a guide for Singapore event vendors

By Karts C · Updated 8 November 2026

If you are asking how to get recommended by ChatGPT as a Singapore event vendor, the short answer is to be named, with your prices and group sizes, on the third-party lists and guides that ChatGPT, Claude and Google's AI answers read, and to let their crawlers read your own website. This guide covers which search index each assistant uses, which crawlers to allow, what to publish on your own site, and the popular fixes with no evidence behind them. It is written for owners and marketers at emcee, catering, team building and venue businesses. Every figure has its source and date in the list at the end.

Where do ChatGPT, Claude and Google find businesses to name?

ChatGPT finds businesses by running a web search when a question needs current information, then reading and citing the pages that come back. OpenAI's crawler for this is OAI-SearchBot, and OpenAI says sites that block it will not be shown in ChatGPT search answers. Which index sits behind the search is only partly public. In November 2024 OpenAI's VP of Engineering said Bing is an important one of the services it uses, and a September 2026 analysis of ChatGPT's traffic found results also coming from OpenAI's own index and from outside scraping providers. For a vendor, that means being indexed in Bing and readable by OAI-SearchBot.

Claude answers from what it learned in training unless it decides to search, and when it searches, the results come mainly from Brave Search. Anthropic added Brave Search to its list of subprocessors on 19 March 2025. A June 2026 re-run of about 400 queries by an AI-visibility tool vendor found 79.2% overlap between the pages Claude cited and Brave's top 10 results. The same vendor's tests found Claude searched in 36.6% of prompts overall and in 81% of prompts asking for the "best" of something. So a question like "best corporate emcee in Singapore" usually sends Claude to Brave, and Brave's crawler skips pages that Googlebot is blocked from.

Google's AI Overviews and AI Mode draw on Google's normal search index. Google says there are no extra requirements to appear in them beyond being indexed and eligible to show a snippet, and no special schema, AI text file or new markup is needed. Both run several related searches behind one question, which Google calls query fan-out, so a page can be cited for a sub-question such as price per pax even if it does not rank for the main query. In a dataset of about 4 million AI Overview URLs reported on 2 March 2026, only 38% of cited pages ranked in Google's top 10.

Which crawlers should your website let in?

Your robots.txt file should allow each assistant's search crawler and its live-fetch bot, since a page they cannot read cannot be cited. Each company runs separate bots for search, for fetching a page when a user asks a question, and for training models. OpenAI, Anthropic, Google and Apple each say that blocking their training bot does not affect search visibility, so you can opt out of training and still be cited. OpenAI says robots.txt changes take about 24 hours to apply, and Perplexity says up to 24 hours. The bots to allow, and the training bots you may block, are below.

Check the file your host actually sends out, and compare it with the one you uploaded. Some content delivery services can add block rules for AI crawlers in front of your own robots.txt and serve both as one file, so open your domain followed by /robots.txt in a browser and read the result. If your prices load only after a script runs, or sit inside an image, check that the numbers appear when you view the page source. Google asks site owners to make important content available in textual form, and an assistant cannot quote a rate it cannot read.

Why do third-party lists count for more than your own website?

AI assistants lean on lists written by someone other than the business being recommended. A study of 75,000 AI answers across ChatGPT, Google AI Mode and Perplexity, reported on 24 March 2026, found listicles made up 21.9% of all citations and 40.9% of citations for commercial questions, the kind a planner asks when looking for a caterer for 200 pax. In professional services, 80.9% of the cited listicles were written by third parties and 19.1% were a company's own list with itself on it. A "top 10 emcees" page on your own website is the weaker version of the same thing.

Being named on other sites also tracks AI visibility more closely than website authority does. A December 2025 study of 75,000 brands found the number of web pages mentioning a brand correlated with its visibility in ChatGPT, AI Mode and AI Overviews at 0.66 to 0.71 (Spearman), while Domain Rating correlated at 0.27 to 0.33 and backlinks only weakly. A page counted as a mention even without a link. For an event vendor, the work this points to is getting named on Singapore guides, venue partner pages, association member directories and press coverage, with your name spelled the same way each time.

One Singapore test points the same way. In about 40 prompts run across ChatGPT, Gemini and Perplexity and published on 31 October 2026, a local agency found interior design firms with home-renovation directory profiles and press coverage were recommended ahead of firms with stronger websites but fewer outside mentions. It did not cover event vendors and did not publish its prompts. Lists are changing too: ChatGPT's citations of "best" and "top" pages fell 30% from December 2025 to January 2026, and the lists that kept growing had 10 to 20 options, the current year on the page and a stated method of selection.

Read these figures as a rough guide. Most of the studies in this article are published by companies that sell AI-visibility tools or services, most samples are from the US, and they measure correlation: they show what cited pages have in common and cannot show what caused a citation. No study has looked at Singapore corporate event prompts. The findings still agree with each other across different datasets, and the actions they point to cost a vendor very little to try.

What should you publish on your own website?

Publish the facts a planner filters on, in plain text on a normal web page: price per pax or per hour, minimum and maximum group size, halal status and the certifying body, the areas or venues you cover, what the package includes, and how much notice you need. These are the details people put into prompts, such as a halal buffet for 150 pax in Jurong within a set budget per head. If your rates exist only in a PDF, an image or a WhatsApp reply, an assistant has nothing to match. A July 2026 write-up of a 17.2 million citation analysis recommends HTML pages over PDFs.

Write each service as a short block whose first sentence answers the question on its own. In a study of 15.7 million Google AI Mode citations reported in July 2026, the median quoted passage was 117 words, and 80% of quoted passages put the answer in the first sentence. A first sentence that gives the service, the starting price per pax, the minimum order and the delivery terms can be lifted whole into an answer. A sentence such as "We cater for all occasions" gives an assistant nothing to quote, and it says nothing a buyer can compare.

Use numbers and date them. In an April 2026 study of 602 prompts, pages containing numbers and statistics had 61.55% more of their text reused in AI answers than pages without them. Add a visible line such as "Prices checked 8 November 2026" and change it only when you have re-checked the prices. A July 2025 study of about 17 million cited URLs found AI-cited pages were 25.7% fresher by publish date than pages in ordinary search results. If you add structured data, Google says it must match what the visible page shows.

Keep your business name, address and details the same everywhere

Use one business name, one address or service area, one phone number and one website address on every profile you control. Assistants piece a business together from many pages, and mismatched details make it harder to tell that two pages describe the same vendor. Google says businesses with complete and accurate information are more likely to show up in local search results. A September 2026 study of 3,850 supplier-selection prompts found directory entries with one business per entry and consistent attributes were cited 3.1 times as often as free-text listings. The company behind that study runs the directories it tested.

Start with the name registered with ACRA and decide how your trading name relates to it. If you trade under a shorter brand name, use it the same way on your website, Google Business Profile, LinkedIn page, Facebook page and every directory, and put the registered name in your website footer. Update old listings when you move or change numbers, including the ones you set up years ago and forgot. A Singapore local-citation guide updated in October 2026 warns that name, address and phone details must match exactly, and suggests adding new listings a few at a time.

Do Google reviews affect AI recommendations?

Google's AI answers cite Google Business Profiles often, so your reviews and profile details are part of what they read. In an August 2026 study of 1,355 business locations in the US, UK and Australia, Google Business Profiles made up about 28.6% of all citations in local AI answers and were cited for 94% of the businesses checked. Those were mostly consumer categories such as restaurants, dentists and lawyers, so the share for Singapore corporate event vendors is untested. Google says more reviews and positive ratings can help a business's local ranking, and that there is no way to request or pay for a better one.

Ask for reviews the way Google allows. Google lets businesses share a review link or QR code with customers, and says offering free or discounted goods or services in exchange for reviews is strictly prohibited. Send the link after the event, when the client has just seen your work, and reply to the reviews you get; Google says replies show customers you value their feedback. If you travel to clients and have no premises open to them, a service-area business can show an area in place of an address. Only businesses that meet customers face to face are eligible for a profile.

Reviews on other platforms count as well. A November 2025 study of 129,000 domains found pages present on several review platforms averaged 4.6 to 6.3 ChatGPT citations, against 1.8 for pages absent from review sites. The researchers noted that review and community presence helps smaller sites in particular. For a B2B vendor, the review platforms worth the effort are the ones your clients already check before they book. Never write or buy reviews.

What does not help: llms.txt, FAQ schema and link attributes

An llms.txt file has no confirmed users among the major assistants. Google's John Mueller wrote in June 2025 that no AI system currently uses llms.txt, and Google's Gary Illyes said in July 2025 that Google does not support it and is not planning to. The crawler documentation of OpenAI, Anthropic and Perplexity does not mention it. A Q2 2026 study of 96,418 URLs found no measurable advantage for sites that had one. The file does no harm if you already have it. If someone wants to charge you for writing one, the money buys nothing anyone has measured.

FAQ schema no longer earns anything visible in Google. Google stopped showing FAQ rich results in search from 7 May 2026 and removed its documentation for that markup. A November 2025 study found pages with FAQ schema averaged 3.6 ChatGPT citations against 4.2 for pages without it. Keep a questions section on your pages if clients really ask those questions, because the visible text can still be read and quoted by assistants. The markup behind it is optional, and paying for an FAQ schema package is money spent on a result Google has switched off.

Whether a link to you is followed, nofollow or sponsored barely changes whether AI assistants mention you. A study of 1,000 domains published in October 2025 found followed links and nofollow links correlated with AI mentions almost identically, at 0.504 and 0.509 (Spearman). Assistants read the text of a page, so the part to check is whether your name, category and prices appear on it. Paying extra for a "dofollow" link buys a signal the study found makes no difference, and Google's spam policy treats paid links that pass ranking credit as link spam.

How do you check whether ChatGPT and Claude name you?

There is no fixed position to rank on in ChatGPT, so the useful version of "how to rank on ChatGPT" is how often you are named for the prompts your buyers type. Pick five to ten prompts a client would ask, such as a corporate emcee for a 300 pax dinner and dance in Singapore, and run them once a month in ChatGPT, Claude, Gemini and Perplexity. Record whether you are named, who else is named, and which pages are cited. Expect variation: a July 2026 study of 161,286 prompts found only 3.8% of sources were cited by all four engines for the same prompt.

Two free tools help. Bing Webmaster Tools added an AI Performance report in February 2026 that shows how often your pages are cited in Microsoft Copilot and Bing's AI summaries, with sample grounding queries, the phrases Bing searched when it used your page. Because Claude relies mainly on Brave, searching your main question on search.brave.com shows roughly the results Claude will read. If you would rather pay for tracking, self-serve tools start at US$29 a month for 15 prompts, checked on 8 November 2026. Compare a month of manual checks against that before you buy.

Most of what agencies sell in Singapore as AI SEO, GEO or AEO is on the checklist below. A vendor with an afternoon a month can do the first round alone: open the crawlers, put prices in text, tidy old listings, ask clients for reviews and get named on the lists assistants already cite. Run it for three months and keep the prompt log, then decide whether outside help is worth paying for. If you do hire someone, the log gives you a baseline to check their reports against.

Checklist

Questions vendors ask

Does my website need to rank on Google first to be recommended by ChatGPT?

No. An August 2025 study of about 15,000 prompts found only 8% of ChatGPT's cited pages ranked in Google's top 10 for the same prompt, and in data on about 4 million AI Overview URLs reported in March 2026 only 38% of AI Overview citations came from top-10 pages. Ranking in the assistant's own search counts for more: Bing and OpenAI's own index for ChatGPT, Brave for Claude.

Which crawlers must I allow for ChatGPT and Claude?

For ChatGPT, allow OAI-SearchBot and ChatGPT-User. For Claude, allow Claude-SearchBot and Claude-User. OpenAI and Anthropic say their training crawlers, GPTBot and ClaudeBot, are separate, so you can block those and still appear in search answers. OpenAI says robots.txt changes take about 24 hours to apply.

Does an llms.txt file help my business appear in AI answers?

There is no evidence that it does. Google's John Mueller said in June 2025 that no AI system currently uses llms.txt, the crawler documentation of OpenAI, Anthropic and Perplexity does not mention it, and a Q2 2026 study of 96,418 URLs found no measurable advantage for sites with one.

Do nofollow links from directories still help with AI recommendations?

As far as the data shows, they count about as much as followed links. A study of 1,000 domains published in October 2025 found nofollow links and followed links correlated with AI mentions at almost the same level (0.509 and 0.504, Spearman). What assistants use is the text on the page: your name, category, prices and location.

How long until ChatGPT starts naming my business?

No study has measured it. OpenAI says robots.txt changes reach its crawler in about 24 hours, and one site's server logs in June 2026 showed Bing crawling 40 new articles within 3 to 6 hours of an IndexNow ping. Being crawled is only the first step, so check your prompts monthly for three to six months before judging.

The Shortlist lists Singapore event vendors for free. To be considered, send your details through the submit a listing form.

  1. Overview of OpenAI crawlers, OpenAI developer docs, read 8 Nov 2026
  2. OpenAI VP of Engineering on Bing and ChatGPT search, Search Engine Roundtable, 4 Nov 2024
  3. ChatGPT built its own search index, Peec AI research (reverse-engineering of ChatGPT traffic, 21 May to 21 Jul 2026), Sep 2026
  4. Anthropic crawler documentation (ClaudeBot, Claude-SearchBot, Claude-User), Anthropic Help Center, read 8 Nov 2026
  5. Claude web search, Brave Search and Anthropic's subprocessor list, Xponent21, 2026
  6. Profound re-run of about 400 queries, 79.2% Claude and Brave top-10 overlap, as reported by Xponent21, Jun 2026
  7. Claude visibility and Brave Search rankings (Profound session findings: 36.6% search rate, 81% for best prompts), Search Engine Land, mid-2026
  8. AI Overviews and AI Mode guidance for site owners, Google Search Central, read 8 Nov 2026
  9. AI Overview citations from top-ranking pages drop sharply (Ahrefs, 863,000 keywords, about 4M URLs), Search Engine Journal, 2 Mar 2026
  10. AI search overlap with Google top 10 (about 15,000 prompts), Ahrefs, 11 Aug 2025
  11. Perplexity crawlers (PerplexityBot, Perplexity-User), Perplexity docs, read 8 Nov 2026
  12. About Applebot and Applebot-Extended, Apple Support, read 8 Nov 2026
  13. Managed robots.txt for AI crawlers, Cloudflare docs, read 8 Nov 2026
  14. AI citations favour listicles, articles and product pages (Wix Studio AI Search Lab, 75,000 answers, 1M+ citations), Search Engine Land, 24 Mar 2026
  15. AI brand visibility correlations (75,000 brands), Ahrefs, 12 Dec 2025
  16. Singapore brand recommendation test across ChatGPT, Gemini and Perplexity (about 40 prompts), Hashmeta, 31 Oct 2026
  17. The listicle window is closing in AI search (2M+ citations, Nov 2025 to Feb 2026), Seer Interactive, 25 Feb 2026
  18. Write-up of Yext's Q4 2025 analysis of 17.2M AI citations, Radiant Elephant, 24 Jul 2026
  19. Pillarbase study of 15.7M Google AI Mode citations, summary by Relevant Audience, Jul 2026
  20. Citation absorption study (602 prompts, 18,151 pages), arXiv 2604.25707 v2, 29 Apr 2026
  21. Do AI assistants prefer to cite fresh content? (about 17M cited URLs), Ahrefs, 28 Jul 2025
  22. Directories and AI citations on supplier-selection prompts (Citations.press, 3,850 prompts), summary by DerivateX, 21 Sep 2026
  23. Local SEO citations in Singapore, marketingagency.sg, updated 26 Oct 2026
  24. How to improve your local ranking on Google, Google Business Profile Help, read 8 Nov 2026
  25. Tips to get more Google reviews, Google Business Profile Help, read 8 Nov 2026
  26. Business Profile eligibility guidelines, Google Business Profile Help, read 8 Nov 2026
  27. AI directory sources study (1,355 locations, 1.97M citations), BrightLocal, 26 Aug 2026
  28. Top factors influencing ChatGPT citations (SE Ranking, 129,000 domains), Search Engine Journal, 26 Nov 2025
  29. John Mueller: no AI system currently uses llms.txt, Search Engine Roundtable, 18 Jun 2025
  30. Gary Illyes: normal SEO works for AI Overviews, llms.txt won't be used, Search Engine Land, Jul 2025
  31. SISTRIX study of 96,418 URLs (Q2 2026), summary by Vizup, 29 Oct 2026
  32. FAQ structured data retirement notice, Google Search Central, 2026
  33. Backlinks and AI search study (1,000 domains), Semrush, Oct 2025
  34. How AI really weighs your links (Kevin Indig and Semrush), Search Engine Journal, 1 Oct 2025
  35. Spam policies for Google web search (link spam), Google Search Central, read 8 Nov 2026
  36. AI citation source overlap study (161,286 prompts), Writesonic, 22 Jul 2026
  37. Bing Webmaster Tools AI Performance report, Search Engine Roundtable, 11 Feb 2026
  38. Otterly.AI pricing (Lite US$29 a month, 15 prompts), checked 8 Nov 2026
  39. IndexNow speed test (40 articles, server logs), Tygart Media, 22 Jun 2026

More on AI visibility for Singapore vendors · Submit a listing · Advertise