A crane rental operator in Johor came to us frustrated. They had been running Google Ads for years, were known in the regional industrial circuit, and their site ranked decently for the obvious keywords. None of that mattered when their largest prospects had started asking ChatGPT before opening Google.
Their question to us: when a project manager at a port logistics company in Pasir Gudang asks an AI engine “I need a 200-tonne mobile crane for two weeks, who can deliver fast?”, does our brand come up?
It did not. On any of them.
The brief
Two goals.
- Get the brand cited by AI engines for high-intent buyer queries in the southern Malaysia industrial market.
- Do it without abandoning the existing Google Ads pipeline that was still producing leads.
The challenge
Niche B2B services have a specific AI visibility problem. The volume of content written about “200-tonne crawler crane rental Johor” is small. AI engines therefore lean heavily on whatever credible sources they can find, usually directories, association sites, and a handful of operator websites that have written specifically about the buyer’s question.
If you are not in that small set of sources, you are invisible. Even if you have been operating for 20 years.
What we did
1. Mapped 18 real buyer prompts
We sat with the sales team and pulled actual buyer enquiries from the past 12 months. Then we wrote out, in the buyer’s language, the kinds of questions those buyers would have asked an AI engine.
Examples:
- “Where can I rent a 200-tonne mobile crane in Pasir Gudang for a two-week shutdown?”
- “What is the typical lead time for booking a 500-tonne crawler crane in southern Malaysia?”
- “Which crane operators in Johor handle port-side lifts at major terminals?”
Specific. Long. Project-manager voice. Not “best crane rental Johor”, that’s the SEO version of the same idea, but it is not what an AI engine sees as a real buyer prompt.
2. Rebuilt location and capability pages around the prompts
Each prompt got an answer on a real page on the site. Not a generic services page. A page that named the project type, the equipment, the response time, and the area served, in the same language the buyer would have used.
Schema markup was added: LocalBusiness with full address and area served, Service schema for each crane class, FAQPage for the prompt-style questions.
3. Earned third-party mentions
We identified the directories, association sites, and trade publications that AI engines pull from for industrial service queries in Malaysia. The operator already had relationships with several. We helped them upgrade their listings, file association applications they had been putting off, and contribute one bylined article to a regional logistics publication.
4. Tracked AI visibility weekly
We ran the 18 prompts on ChatGPT, Perplexity, Gemini, Copilot, and Google AI Overviews every week. Built a simple scoring sheet so the operator could see month-on-month progress.
Results
- First AI citation appeared on Perplexity within 38 days.
- By day 90, the brand was being mentioned on 4 of 5 AI engines for the highest-intent prompts.
- Inbound enquiries that referenced “we asked ChatGPT” went from zero to a steady 2-3 per week.
- Existing Google Ads pipeline was untouched, both channels now run in parallel.
What we learned
For niche B2B in Malaysia, AI SEO is a faster game than traditional SEO. Traditional SEO compounds over months because Google needs to crawl, evaluate, and re-rank. AI engines update their training and retrieval more frequently, and they reward specificity heavily.
If you can write the answer to a buyer’s exact question in their exact language, and you have one or two credible third-party signals, you will start showing up faster than a comparably sized SEO play could deliver.