How a Malaysian recruitment and training firm got found in AI search: the Carriera case study

Carriera is a Malaysian recruitment and corporate-training group. It is an ordinary services business in a crowded category, and it is not our own company — which is the point of publishing this alongside the DNE case study. In the 30 days to 27 August 2026 its site drew 768 visitors against 230 in the previous 30 days, Google showed it 18,730 times against 6,240. These figures measure website traffic and Google visibility; they do not establish what share came from AI assistants.

Editorial correction: 14 September 2026. Measurement snapshot: 28 August 2026. Figures pulled from Carriera's Google Analytics 4 and Google Search Console on that date.

Key takeaways

• In the 30 days to 27 August 2026, carrieragroup.com.my drew 768 visitors across 970 sessions. In the preceding 30 days it drew 230 visitors across 302 sessions. That is +234% visitors in one month.
• Google Search Console recorded 238 clicks from 18,730 impressions, against 47 clicks from 6,240 impressions the month before — impressions tripled and clicks rose roughly five-fold.
• We publish no figure for how much of this came from AI assistants, because the analytics cannot produce an honest one. Why, and what we would need to run instead, is set out below.
• Measurement on this property began 19 June 2026, so June is a partial month. July, the first complete month, recorded 265 visitors; August recorded 725 through the 27th.
• This is an arm's-length client. Our other published case study, DNE Logistics, is the founder's own company. Carriera is not. If you have been waiting to see whether the method works on a business we do not own, this is that page.
• The honest limits are stated in full below: the enquiry numbers are small, one conversion path is not tracked at all, and average ranking position slipped while reach grew.

On this page8 sections

Who is Carriera?

Carriera operates as two registered Malaysian entities: Agensi Pekerjaan Carriera Talent Resources Sdn Bhd, which holds Ministry of Human Resources private employment agency licence JTKSM 615, and Carriera Academy Sdn Bhd, an HRD Corp approved training provider. The first places white-collar staff; the second runs claimable corporate training.

That combination is exactly the kind of business AI search is difficult for. Both halves live in categories where buyers ask broad, comparative questions — which agencies recruit in this sector, which training is HRD Corp claimable, how the levy works — and where dozens of competitors publish near-identical marketing pages. Being the answer means being the clearest and most verifiable source on those questions, not the loudest.

What was built?

The live site is 78 pages, up from 23 when the work started. The shape of it matters more than the count:

1. Answer-shaped pages for the questions buyers actually ask. Sector-by-sector recruitment pages, HRD Corp training-theme pages, a salary guide, a recruitment-process explainer, and comparison pages — each written to answer one real question completely rather than to rank for one phrase.
2. Two working calculators, for the HRD Corp levy and for cost of hire. A tool that returns a real number is the kind of page other people link to and assistants quote.
3. A verifiable entity layer. Both SSM-registered company names, the JTKSM 615 licence, the HRD Corp registration, real address, phone and hours, wired into structured data and stated consistently in text. An assistant can confirm who Carriera is without guessing.
4. Corroborating profiles that agree with each other. The company LinkedIn, Facebook, Google Business Profile and YouTube all reference the same entity and the same site, so independent sources tell the same story.
5. A live jobs board and training catalogue running off Carriera's own backend, so what a visitor sees is the real current vacancy and course list, not a stale brochure.
6. A PDPA privacy notice, because the site collects names, emails, phone numbers and CVs, and it previously had no privacy or terms page at all.

Everything above is verifiable by opening the site. We have deliberately not attributed a share of the growth to any single item on that list, because the analytics cannot separate them.

What happened to the traffic?

By complete month, the first full month of measurement was July at 265 visitors; August reached 725 by the 27th, with four days still to run.

Search Console tells the same story from Google's side:

Impressions roughly tripled. That is the number worth watching: it means Google is now putting Carriera in front of people for roughly three times as many searches as it was a month earlier.

Is AI actually sending people?

We cannot tell you from the analytics, and we are not going to guess.

Google Analytics does report an "AI Assistant" channel, and an earlier version of this page quoted it. We have stopped. Assistants strip or omit the referrer inconsistently, so an unknown and variable share of AI-prompted visits arrives labelled Direct, or as a brand search after the person looked the name up. Whatever that channel reports is the residue of whichever assistants happened to pass a referrer, not a measure of AI's contribution — and the breakdown by engine is worse, because it mostly measures how each engine handles referrers rather than how often it recommends the business. Framing it as a floor does not rescue it. A caveat attached to a misleading number still leaves the number in the reader's head, and this page previously used that figure to compare Carriera against DNE, which was comparing two artifacts.

The gap we have not closed yet. For our own company we ran multi-engine capture sets — putting realistic buyer questions to each assistant, without naming the business, and recording what came back. That is the evidence that can actually support an AI-visibility claim, and it is in the DNE case study. We have not yet run the equivalent set on Carriera's categories. When we do, the results go here, including the questions where Carriera does not appear.

What the record does show, without needing to attribute anything: search reach tripled in impressions and clicks rose roughly five-fold in a month, visitors went from 230 to 768, and no paid channel was involved. If you want the reasoning behind refusing to publish an AI share, it is in our guide to choosing an AEO agency without getting sold a story.

What this case study does not show

Four honest limits, because a case study that only flatters is not evidence.

The enquiry numbers are small. In the same 30 days the WhatsApp button was tapped 16 times by 11 people, and the email link twice. On a site with 768 visitors that is a low rate, and we are not going to dress it up.

One conversion path is not measured at all. Job applications, contact forms and training registrations post to Carriera's own backend, and those submissions are not currently tracked as events in Google Analytics. For a recruitment business those are the conversions that matter most. So the real enquiry volume is higher than 16 — we simply cannot tell you by how much, and the fix is instrumentation, which is now the next job.

Average position went down, not up. From 8.3 to 9.8. That is what usually happens when a site starts ranking for far more queries than before: the new ones enter lower and pull the average down while total reach triples. It is a good trade, but it is a trade, and a report that hid it would be misleading.

Engagement rate fell from 43% to 33%. Same cause — a much wider mix of visitors, many arriving on informational pages. Worth watching rather than worth panicking about.

And one month is one month. Two 30-day windows are enough to show a direction, not enough to prove a trend. We will update this page as the record lengthens, rather than leaving a flattering snapshot up forever.

What would we tell another Malaysian business?

The order of work is the lesson, and it is the same AI-search visibility programme we run for every client.

• Make the verifiable facts machine-readable first. Licence numbers, registered entity names, address, hours. Carriera had two real SSM entities and a JTKSM licence and none of it was stated where a machine could read it. That is the cheapest work with the largest effect on whether an assistant will name you at all.
• Answer whole questions, not keywords. A page that fully answers "is this training HRD Corp claimable and how do I claim it" is quotable. A page optimised for "HRD Corp training" is not.
• Build one thing that returns a real answer. A calculator, a checker, a comparison table. It earns links and it gets quoted.
• Instrument the conversions before you need the report. Carriera's biggest measurement gap exists because the forms were live before the tracking was. Do it the other way round.

If you want to know where your own business currently stands, the five-minute version you can run yourself is in how to check if ChatGPT recommends your business in Malaysia. Or book a call and we will ask the assistants your customers' real questions with you on the line, and show you whether the answer names you or a competitor. The rest of our published work is on our work.

When your customer asks AI who to trust, you want one name to come back. Make it yours. Be the answer.

Sources

• Carriera, Google Analytics 4 property 542243236 (first-party data), 30 days to 27 August 2026 and monthly totals from 19 June 2026, Asia/Kuala_Lumpur time zone. Pulled 28 August 2026. Held privately by Carriera, not a public URL.
• Carriera, Google Search Console property https://carrieragroup.com.my/, 30 days to 27 August 2026 against the preceding 30 days. Pulled 28 August 2026.
• Entity and licence details verified on carrieragroup.com.my on 28 August 2026: Agensi Pekerjaan Carriera Talent Resources Sdn Bhd (JTKSM 615) and Carriera Academy Sdn Bhd (HRD Corp approved training provider).
• Page count of 78 verified from the live sitemap on 28 August 2026.