This case study documents a nine-month engagement with a Singapore IT managed services provider that wanted to show up when business owners ask AI assistants for help. Choosing an IT provider is exactly the kind of decision people now research in ChatGPT, Perplexity and Google's AI Overviews: "best IT support company for a 20-person office in Singapore", "how much does managed IT cost", "who can help us after a ransomware attack". AI tools answer with a handful of named firms. If yours is not one of them, the buyer may never reach a search results page.
We started by testing 60 prompts that the firm's prospects might ask. The firm was named in 3. It appeared in Google AI Overviews for 4 of 40 tracked queries. Its business details were inconsistent across its website, Business Profile, LinkedIn and directories, and its About page said little beyond a mission statement. Bing, which ChatGPT search relies on, had indexed only 23 of its 118 pages. AI referral traffic was 12 sessions a month, and no leads had ever been traced to it. Classic SEO was modest too: 2,380 organic visitors a month and 46 page 1 keywords.
We built a fixed set of 60 prompts from sales calls and enquiry emails, covering "best" and comparison questions, cost questions and problem questions. Each month we ran every prompt three times in ChatGPT and in Perplexity, and counted the firm as cited when it was named or linked in at least two of three runs on either tool. We tracked AI Overviews separately on 40 Google queries. In GA4 we created an AI referrals channel group for chatgpt.com, perplexity.ai, gemini.google.com and copilot.microsoft.com, and added "AI assistant" to the "How did you hear about us" field on the enquiry form.
AI answers change from one run to the next, so a single screenshot proves nothing. A fixed prompt set, repeated runs and a clear counting rule turned AI visibility into a number the client could track month to month, and the GA4 channel group showed whether that visibility turned into visits and leads.
We made the firm's core facts identical everywhere: registered name, address, phone number, year founded, services and sectors served. That covered the website, Google Business Profile, Bing Places, LinkedIn and 19 directories. We rewrote the About page around specific, checkable facts, such as team size, engineers' certifications and the industries the firm supports, and added author pages for the engineers writing its guides. Organization and LocalBusiness schema now link all of these profiles together.
AI tools describe a business using what they can find and confirm across many sources. When the facts disagree, they either leave the business out or describe it wrongly. Consistent facts and a clear About page made it easy for them to say who the firm is and what it does.
We verified the site in Bing Webmaster Tools, submitted the sitemap and fixed the issues blocking indexing. Pages indexed in Bing rose from 23 to 114 of 118. We also found that the site's firewall was blocking the crawlers ChatGPT and Perplexity use for search, and updated its rules to allow them.
ChatGPT's search draws heavily on Bing's index. A page Bing has not indexed, or a crawler the site blocks, cannot be cited however good the content is. This was the least visible fix in the project and one of the most important.
We rewrote the 12 service pages and published 24 guides, each built around a real buyer question. Every section opens with a direct two or three sentence answer, then the detail. Cost pages give ranges in S$, such as S$1,200 to S$2,400 a month for managed support for a 20-person office, with what changes the price. Guides use comparison tables and checklists, name the engineer who wrote them and show a last-updated date.
AI tools quote passages that answer a question clearly and on their own. Pages that bury the answer under an introduction are rarely cited. The same structure also won featured snippets and AI Overviews in Google, which is where most of the classic organic growth came from.
For "best IT support" style prompts, AI tools lean on what other sites say. We pitched the firm to the editors of 14 Singapore business roundups and comparison articles with facts they could check, completed its profiles on two SME associations, and secured guest articles in three business publications. The firm asked every client at the end of a project for a Google review, taking reviews from 38 to 112.
AI assistants rarely recommend a firm on its own say-so. They repeat what independent sources say. Being named in roundups, association listings and reviews is what moved the firm into answers for the "best" and "recommended" prompts.
There is a lot of noise about AI search. These are the things we saw make a measurable difference on this project, and one that did not.
Once the firm's name, services and details matched everywhere, AI tools started describing it correctly. Before that, two tools listed services it does not offer.
Every "best IT support" prompt the firm entered came after it appeared in roundups, association listings or reviews. Its own website alone did not get it recommended.
Cost and how-to prompts started citing the firm once its pages answered the question in the first two or three sentences, with real S$ figures.
We added an llms.txt file in month 2 and saw no measurable change from it. The gains came from the combination of indexing, clear facts, quotable content and independent mentions, built over months.
AI search visibility is built on the same foundations as SEO, plus a few extra checks. If you run a B2B firm in Singapore, make sure Bing indexes your site, AI crawlers are not blocked and your business facts match everywhere. Without these, nothing else can work.
You cannot get recommended by AI on your own say-so. For "best" and "recommended" questions, AI tools repeat what independent sources say about you. Reviews, roundups and association listings matter more here than anything on your own site.
Measure it, or you are guessing. Pick a fixed set of prompts your buyers would ask, run them every month in the same way, and track AI referral sessions and leads in GA4. A single good answer in a screenshot tells you very little.
No one can guarantee it, because ChatGPT decides its own answers. What we can do is make your business easy to find, understand and trust: indexed in Bing, consistent facts across the web, pages that answer buyer questions clearly, and mentions and reviews on independent sites. On this project, that took the firm from 3 to 27 of 60 tracked prompts.
We track a fixed set of prompts your buyers would ask, run each one several times a month in ChatGPT and Perplexity, and count how often you are named or linked. We track Google AI Overviews on a separate set of queries. In GA4, a channel group shows sessions and leads from chatgpt.com, perplexity.ai and other AI tools.
Yes. AI tools with web search pull from search indexes and cite pages that rank and answer clearly. On this project, the same work that won AI citations also grew classic organic traffic by 169%. AI search adds checks such as Bing indexing and crawler access, but it does not replace SEO.
Fixes to indexing and crawler access can show within weeks. Being recommended for competitive "best" questions usually takes several months, because it depends on mentions and reviews building up on other sites. This firm went from 3 to 8 cited prompts by month 3 and 27 by month 9.
Start with a free SEO and AI search audit. We test the prompts your buyers are likely to ask, check your Bing indexing, AI crawler access and business facts across the web, and review the pages AI tools could quote, then show you a realistic path to being cited before you commit to anything.
Singapore’s specialist SEO agency for SMEs. We rank your business on Google — and only Google. No distractions, just results.
© 2026 Singapore SEO Agency. All rights reserved.