BlackRevo blog · AI trends
AI at work in Malaysia: what 2026 looks like, how we use it every day, and what we expect from you
Yap Pat Yih, Co-founder and General Manager · 2026-10-05 · 9 min read
93% of Malaysian employees use generative AI at work; 12% were trained for it. What the 2026 surveys say, how we use AI daily, what we never let it do, and what AI skills mean.
Most of the Malaysian workforce is already using AI at work, and most of it was never taught how. EY's 2025 Work Reimagined Survey, published in January, put generative AI use at 93% of employees and sufficient training at 12%. Microsoft's June report found a quarter of Malaysian knowledge workers are what it calls Frontier Professionals, ahead of the world, while only a third say their leaders are clearly aligned on AI. The gap is not between people who use AI and people who do not. It is between companies that have decided what AI is for and companies that have not. We run a small SEO and web company where AI is in the daily work of every role, so this is what we decided, what the data says, and what we ask of the people we hire.
TL;DR
- 93% of Malaysian employees use generative AI at work; 12% get sufficient training (EY 2025 Work Reimagined Survey).
- 24% of Malaysian workers are Frontier Professionals, against 16% globally; 92% of AI users treat AI output as a starting point (Microsoft, June 2026).
- 68% of active users say their workload went up, not down (EY).
- 620,000 Malaysian jobs, about 18% of those in the ten sectors studied, are expected to be highly impacted within three to five years (TalentCorp).
- At BlackRevo, AI drafts and a person decides. There are three things it never does for us.
- "AI skills" in our hiring means judgement about output, not a list of tools.
What the 2026 numbers actually say
Everyone is using it. EY's Malaysian findings, drawn from 300 employees and 30 employers, show 93% using generative AI at work, with daily use nine percentage points above the global average. 81% of the active users report significant time savings and 76% report improved work performance. Then the sentence that matters: 68% say their workload has increased. The time saved went back into more work, not less, and only 12% receive sufficient AI training to get the full benefit.
Malaysia's best users are ahead of the world. Microsoft's 2026 Work Trend Index surveyed 2,000 knowledge workers in Malaysia and found 24% are Frontier Professionals, using AI to analyse information, solve problems, evaluate and think creatively, against 16% globally. Two habits separate them from everyone else: 57% pause before starting work to decide what a person should do and what AI should, against 39%, and 42% deliberately do some work without AI to keep their skills sharp, against 33%. Across all Malaysian AI users, 92% treat what AI produces as a starting point, not the final answer.
Companies are behind their staff. Only 32% of Malaysian AI users in the same study say their leadership is clearly and consistently aligned on AI, and 19% say they are rewarded for reinventing how work gets done when it does not produce results straight away. People are moving faster than the organisations paying them.
The exposure is real. TalentCorp's impact study, reported by The Edge, found that around 18% of the 3.5 million employed in the ten sectors it examined, about 620,000 jobs, are expected to be highly impacted within three to five years, meaning at risk of convergence or displacement unless people reskill. The ten include wholesale and retail, food manufacturing and services, and global business services (TalentCorp's study page). Digital marketing is not one of the ten, and I would not take comfort from that; the work of an SEO executive in 2023 and the work in 2026 already do not look alike.
How we use AI, every day
AI is in the daily work of every role here, and it got there because a small team has to spend its time on clients instead of paperwork. The honest list, without the tool names, because the tools change every quarter and the rules do not:
- Research first drafts. Keyword clusters, competitor page summaries, first outlines. A person checks the sources before anything is used, because the thing AI produces most fluently is a confident number with no page behind it.
- The boring half of writing. Meta descriptions, alt text, the tenth variant of an ad line, the first pass of a translation into 中文 that a native speaker then rewrites for how people actually talk.
- Reading, not deciding. Summaries of long client documents, call notes, a 40-page report reduced to the questions we need to ask. The decision on what to do with it stays with the person whose name is on the account.
- Our own systems. We built our HR and hiring system in-house, and the way we built it is the way we use AI generally: it takes the paperwork off people so that the hiring conversation, the leave decision and the performance review are done by a human with more time.
That is it. It is not a transformation programme. It is a set of habits that save time every week and are re-examined when they stop doing so.
The three things we never let it do
Client-facing work goes out under a name, not a model. Nothing reaches a client, a candidate or a search engine that a named person has not read and is not willing to defend. If it is wrong, the person is accountable, which is the only reason anyone reads carefully.
Decisions about people. Who we hire, what someone is paid, whether a review is good or bad: these are made by people who have met the person. AI can help read an application; it does not decide on one.
Confidential data. Client pricing, financial data, client names and anything else not already public stay confidential, whatever tool is in front of you. That is a handbook rule, not a suggestion, and it is the question most candidates never think to ask about.
The 92% of Malaysian AI users who treat AI output as a starting point have the first rule right. The second and third are where companies get into trouble, and mostly they get there by never writing the rule down.
What "AI skills" means when we hire
Search for AI skills in Malaysia this year and most of what comes back is a course to buy. Here is what I actually look for, and none of it is a certificate.
- You can tell good output from bad. Show me a piece of AI-drafted work you rejected, and why. That answer says more than any prompt library.
- You know what to do yourself. Microsoft's Frontier Professionals pause to decide which parts of a job a person should do. In an interview I ask the same question about a real task.
- You still have the underlying skill. If you cannot write a paragraph or read a keyword report without help, AI makes you faster at being wrong. 42% of Malaysia's most advanced AI users practise without it on purpose.
- You can say what you did. "I used AI" is not an answer to "walk me through this project". "I had it draft twelve headlines, killed ten, rewrote two, and this is the one that ran" is.
- You know where the line is. A candidate who asks what they may and may not paste into a public tool has already understood something most companies have not.
We do not run an AI test in the interview. We ask people to walk through something they made, and the AI part comes out on its own, or it does not. Our post on answering "tell me about yourself" covers how to talk about a project so that your judgement shows.
If you are worried about your job
Three of the numbers above are worth holding together. 620,000 jobs are expected to be highly impacted. 93% of employees already use the tool behind much of that change. 12% were trained enough to get the full benefit. The risk is not that AI arrives; it is here. The risk is being in the 88% who never got taught what it is for, in a company that has not decided either. You can fix half of that yourself in a month, by using the tools on real work and keeping notes on what they get wrong. The other half is choosing an employer that has written its rules down. Ask about it in the interview. The answer will tell you which kind of company you are talking to, and the open roles at jobs.blackrevo.com.my will tell you ours. EY also found that one in four Malaysian employees still plan to leave, driven partly by wanting AI exposure; what that looks like for young workers is in our job hopping post, and who we are is in the night we joined the 100 Most Influential Young Entrepreneurs.
FAQ
How many Malaysians use AI at work?
93% of the Malaysian employees in EY's 2025 Work Reimagined Survey use generative AI at work, and only 12% receive sufficient AI training. Microsoft's 2026 Work Trend Index found 24% of Malaysian workers are Frontier Professionals, the most advanced AI users, against 16% globally.
Is it okay to use ChatGPT at work?
It depends on your employer's rule, and if there is no rule, ask for one. The safe default: never paste client data, personal data, pricing or anything not already public into a public AI tool, treat every output as a draft, and be able to say which parts you checked.
Will AI replace my job in Malaysia?
TalentCorp's impact study found around 620,000 jobs, about 18% of those in the ten sectors it studied, are expected to be highly impacted within three to five years, meaning at risk of convergence or displacement unless people reskill. The people most at risk are the ones who neither learn the tools nor keep the skill the tools are drafting for.
What AI skills do employers want in 2026?
Judgement about output, the habit of deciding what a person should still do, the underlying skill so you can tell when the AI is wrong, the ability to describe your own part in a piece of work, and knowing what data must never go into a public tool. Certificates are optional; examples are not.
How does BlackRevo use AI?
AI is in the daily work of every role, for research first drafts, routine writing, summaries and the paperwork our in-house HR and hiring system takes off people. A named person checks sources and signs off anything that goes to a client, decisions about people are made by people, and confidential information stays confidential.
Do I need an AI course to get hired?
Not at BlackRevo. Bring one piece of work where AI did part of it, and be able to say which part, what you rejected and why. That is the interview.
Yap Pat Yih: Yap Pat Yih is Co-founder and General Manager of BlackRevo and looks after the team, the hiring and the HR system they run on.