How Malaysian professionals are using AI at work, which tools stuck, what employers allow, and the honest discussion about which jobs are actually at risk.
I'm a marketing manager. My last math class was Form 5. I now work with ML outputs daily and understand enough to ask the right questions. Here's how I got there without a maths degree. The honest prerequisite: You need…
I've been in consulting for 3 years and talk to Malaysian companies about technology regularly. Here's my honest read on where AI actually is versus where the headlines suggest. Genuine adoption: - Customer service chat…
The baseline for professional competence is rising. Here's what every professional needs to be functional in, regardless of sector. Data literacy: Understanding what data you're looking at, spotting when a number doesn'…
After a year of trying almost every AI tool that gets posted about, here's what stayed in my actual workflow. Claude / ChatGPT for drafting: First drafts of documents, emails, proposals. Not for final copy, for breaking…
Used all three seriously for coding tasks. Here's where each actually wins. GitHub Copilot: Best for inline completion and boilerplate generation in the editor. The contextual awareness of your actual codebase is better…
Used it for 6 months. Saves me about 30min/day on boilerplate and documentation. At the USD/MYR rate it is RM45/month. If your hourly rate is above RM30, it pays for itself easily.
My team uses: ChatGPT for drafting, Copilot for code, Midjourney for design concepts, Whisper for meeting transcripts. We have saved roughly 2-3 hours per person per week across the team.
Started with fast.ai top-down approach. Then went back to statistics fundamentals on Khan Academy. Kaggle competitions for practice. Took 18 months to feel competent. Hired as junior ML engineer after that.
It is excellent for explaining concepts, generating quiz questions, and debugging logic. It is terrible for factual accuracy on niche topics. Treat it like a smart study partner who sometimes makes things up.
Use AI for: customising your cover letter for each role, practicing behavioral interview questions, and summarizing company annual reports. Just ensure you still add your personal voice to the results.
For drafting, summarising, brainstorming, always okay if output is verified. For client deliverables, depends on company policy and client agreement. Transparency matters more than the tool itself.
Roles most at risk: data entry, basic content writing, simple code reviews, customer service tier 1. Roles growing: AI prompt engineering, AI auditing, hybrid data+domain roles. Reskill while you still can.
Two conversations run through these threads and they rarely agree. The first is practical: which AI tools people have folded into real work, what their employer's policy allows, and where the output still needs checking line by line. The second is the anxious one about which roles are genuinely exposed in Malaysia, written by people in copywriting, junior development, design and customer support who are watching hiring in their field change. Also here: candidates using AI to write applications and recruiters describing how obvious it is, and threads about learning AI skills without a technical background. Opinions run in both directions, which makes this a better read than most coverage of the subject.
Related topics: Kuala Lumpur, Salary Negotiation, Productivity.