What Is Generative AI and How Malaysian Companies Are Using It

Every time I start a training, I ask the room a simple question:

“Can anyone explain what generative AI actually is?”

Silence. A few nervous laughs. Someone will finally say, “ChatGPT?” like it’s a guess on a quiz show.

These are smart people — government officers, bank executives, HR managers. They’ve been hearing about AI for two years straight. In meetings, in the news, in WhatsApp groups forwarded by their nieces. But when you ask them to explain it, they freeze.

“Tuan Ali, semua orang cakap pasal AI. Tapi saya masih tak faham apa benda ni sebenarnya.”

Fair question. Most explanations are either too technical or too vague. Let me explain it the way I do at the start of training — plain language, and tied to real work.

AI isn’t new. For years, companies have used AI for specific tasks:

• filtering spam
• detecting fraud
• recognising faces
• predicting customer churn
• routing calls

That type of AI is typically trained to do one narrow job and one job only. It learns patterns from large datasets, then makes predictions or classifications.

Useful, but limited. Most people didn’t feel it day-to-day because it lived inside systems.

Generative AI is different because it shows up directly in front of the user.

Generative AI doesn’t just predict a label like “fraud” or “not fraud.” It generates output: text, summaries, drafts, ideas, even code.

Here’s how I explain it, in the most layman way possible.

Imagine someone who’s been writing well since school. A’s for English and BM. Then university. Then 20 years of writing experience. That person has absorbed patterns: how arguments are built, how tone changes, how formal writing differs from casual writing.

If you ask her to write a story about your cat, she can produce something in minutes.

Not because she’s copying a story from somewhere.

Because she has learned the patterns of writing over years — and can produce a new piece of writing on demand.

Generative AI works like that, except it has absorbed far more text than any human could: books, articles, websites, manuals, conversations. So when you ask it to draft an email, or summarise a report, it generates an output based on patterns it has learned.

That’s why it feels like a smart assistant — and why it’s useful immediately.

If you strip away the hype, generative AI is mainly a time compression tool for certain kinds of work:

• emails, memos, minutes
• executive summaries
• proposal drafts
• translating and simplifying language

It doesn’t always give you a perfect first draft. But it often gets you to 70–80% fast — and that saves hours and days.

This is where people feel the value quickly.

A tender document. A policy paper. A meeting transcript. A 50-page report.

You can ask:

• “Extract the key requirements.”
• “What are the deadlines, obligations, and penalties?”
• “Summarise for a busy director.”

Most people don’t struggle because they can’t write. They struggle because they don’t know how to structure.

Generative AI is good at:

• outlines
• slide flow
• pros/cons and counterarguments
• reframing for different audiences

If you give it your SOPs, policy documents, guidelines, or manuals, you can ask questions like you would ask a colleague who has read everything.

That’s why it feels powerful. It turns “searching” into “asking.”

I’m not talking about the hype. I’m talking about what teams are doing quietly, right now, to get work done.

The most common usage is writing and summarising:

• minutes and action items
• drafts of reports and papers
• public communication rewritten into simpler Bahasa Malaysia
• restructuring long documents into a format that management wants

It’s not glamorous. It’s just real work.

Most use is still internal and back-office:

• reading and summarising compliance documents
• internal research and briefing notes
• standardised email templates and customer communication drafts

There’s still caution around anything customer-facing. That’s normal. Compliance will move at compliance speed.

These teams adopt faster because the risks are lower and the outputs are visible:

• onboarding documents
• training materials and quiz questions
• internal announcements rewritten for different tone and clarity

The value is speed, not judgment:

• surfacing key clauses
• highlighting unusual terms
• summarising long contracts and tender documents

It doesn’t replace a lawyer. But it reduces the time spent reading through all the pages to find what matters.

They use it like a leverage tool:

• marketing copy
• FAQ drafts
• social media content
• basic customer replies

It’s not that they suddenly became “AI companies.” It’s that they can now produce work that used to require a bigger team.

Let’s be honest about the limits.

This is the hallucination problem. It generates plausible-sounding content that can be wrong.

If something is factual — names, dates, laws, policies — you verify.

It doesn’t know your department’s tone, your boss’s preferences, your organisation’s history.

You’ll get better results when you provide:

• who the audience is
• what format you need
• what to avoid
• what the organisation cares about

It can suggest. It can draft. It can propose options.

But judgment and accountability stays with you.

• A good finance person becomes faster with AI.
• A person with no finance knowledge becomes dangerously confident with AI.

The tool increases output. It does not guarantee correctness.

After training Malaysian working adults for years, I can tell you this:

Most people can learn the basics in a few hours.

The real barrier is confidence — and the workplace environment around them.

People worry:

• “Can I paste this document?”
• “Will IT flag this?”
• “Will my boss think I’m cheating?”
• “What if I make a mistake?”

So they wait for permission that never comes.

Here’s the better framing:

Start safely. Get small wins. Earn trust. Then expand.

If you’re new, start with low-risk tasks and clean data.

Safe to start with:

• rewriting your own emails
• summarising your own notes
• improving a draft that contains no sensitive data
• converting a long write-up into bullet points

Avoid for now (unless approved):

• confidential documents
• personal data (IC, phone numbers, addresses)
• internal financials
• anything you’d be uncomfortable forwarding to the wrong person

If your organisation has an approved platform (e.g., enterprise tools), use that. If not, keep your starting tasks clean and low-stakes.

If you’re new, start with low-risk tasks and clean data.

Safe to start with:

• rewriting your own emails
• summarising your own notes
• improving a draft that contains no sensitive data
• converting a long write-up into bullet points

Avoid for now (unless approved):

• confidential documents
• personal data (IC, phone numbers, addresses)
• internal financials
• anything you’d be uncomfortable forwarding to the wrong person

If your organisation has an approved platform (e.g., enterprise tools), use that. If not, keep your starting tasks clean and low-stakes.

Take a report you’ve already written (sanitised if necessary). Paste it in. Ask:

• “Summarise this in 5 bullet points for a director.”
• “Rewrite this email to be firm but polite.”
• “Turn this into a one-page briefing note.”
• “What’s unclear or weak in this argument?”

You’ll learn more in ten minutes of doing than from reading ten articles — including this one.

The technology will keep moving. Tools will change. Features will come and go.

But the skill that lasts is simple: knowing how to work with AI to produce better work faster — without losing judgment.

I’m Ali Reza Azmi, Founder and Lead Trainer at Twenty-Four Consulting.

Everything I’ve shared here comes from time spent in real training rooms — working with civil servants, managers, executives, and teams across ministries and companies in Malaysia. These are people doing serious work, under real constraints, trying to understand what generative AI actually means for their roles — beyond the headlines and hype.

At Twenty-Four Consulting, we help organisations build practical capability in generative AI, Canva, and modern digital workflows. Our focus is not on tools alone, but on helping teams apply them responsibly, confidently, and in ways that make their work genuinely easier.

If you’re looking to upskill your team — whether you’re introducing AI for the first time or trying to move from experimentation to everyday use — contact us at Twenty-Four Consulting to discuss training for your organisation.