Pekerja Asing and AI: Same Problem, Different Decade

This already happened once

If you run a business in Malaysia, you've already lived through a version of this.

There's a yong tao fu stall I used to go to back home. Nothing fancy — plastic tables, the uncle making everything by hand, same spot for years. Everyone in the neighbourhood knew the taste. You didn't need to check Google reviews. You just knew.

Then at some point, he hired pekerja asing to handle some of the prep. Smart move — he was getting older, business was good, he needed to scale. Hire foreign workers, delegate the repetitive parts, focus on what matters.

But the regulars noticed.

"Lost the taste already."

"Last time better lah."

Maybe the recipe changed. Maybe it didn't. Maybe the fishball filling was slightly different, or maybe people just felt like it was different because they saw new faces behind the counter. It doesn't matter which one is true. The result was the same: people started going less.

This is exactly what's happening now with AI — just faster.

Same problem, different decade

Hiring pekerja asing was Malaysia's first automation wave. Restaurants, factories, construction, plantations — every industry went through the same calculation: we can scale faster and cheaper by delegating work to someone else. And every industry discovered the same tension: some things you can delegate, and some things you can't.

The restaurants that hired foreign cooks for the prep — washing, cutting, portioning — kept their regulars. The ones that handed over the wok lost something. Not always the recipe. Sometimes just the trust.

AI is the same calculation, compressed into months instead of decades. It can write your proposals, answer your customers, generate your reports, manage your scheduling, draft your emails, reconcile your invoices. The question isn't whether you can automate something. It's whether you should.

Because here's the thing nobody talks about enough: automation changes how your customer perceives you, even when it doesn't change your output.

A consulting firm starts using AI to draft strategy recommendations. The recommendations are good — maybe even better than before. But a client finds out, and suddenly the RM20,000/month retainer feels different. "I'm paying for their thinking, not ChatGPT's." The work didn't change. The trust did.

A design agency uses AI to generate first drafts. Faster turnaround, same quality after human review. But the client sees an AI watermark in the metadata of a file. Now every deliverable is under suspicion. "Did a person actually work on this?"

This is the yong tao fu problem. The taste might be the same. But once the perception shifts, it's very hard to get it back.

The handmade parts

Every business has what I call "handmade" parts — the things that define the experience. The reason customers chose you over the other options.

For the yong tao fu uncle, it was the food itself. The hand-stuffed tofu, the broth he'd been refining for years.

For a lawyer, it might be the way they explain a complex situation in plain language during a call.

For a designer, it's the thinking behind a layout choice — not the layout itself.

For a property agent, it's the honest opinion about whether a unit is actually worth it — the message that says "I wouldn't buy this one" when a lesser agent would just push the sale.

These aren't inefficiencies waiting to be optimised. They are the product. They're the reason your customer tells their friend about you instead of the ten other options they could've picked.

Automate these, and you might save three hours a week. But you lose the thing that made those three hours worth paying for.

The commodity parts

Then there are the parts nobody notices. The parts your customer never sees, never thinks about, and would never ask about.

Invoice generation. Appointment reminders. Data entry from forms into spreadsheets. Pulling numbers into a monthly report. Sorting emails into categories. Formatting documents.

Nobody has ever chosen a law firm because their invoices arrive beautifully. Nobody picks a workshop because the appointment reminder came through fast. These are invisible processes — necessary, but not differentiating.

This is where automation is pure upside. Every hour you save on commodity work is an hour you can reinvest in the handmade parts. The strategy call gets more prep time. The proposal gets a more thoughtful angle. The customer gets a reply that actually sounds like you, because you wrote it instead of rushing through it between admin tasks.

The optics problem

Here's the part most people miss.

Even when automation doesn't change the output at all, the knowledge that something was automated changes how people feel about it.

Think about handwritten birthday cards versus printed ones. The message could be identical. But the handwritten one means more, because someone took the time. The value isn't just in the words — it's in the effort behind them.

Now scale that to business. A client receives a weekly check-in email from their account manager. Thoughtful, specific, clearly aware of what's happening in their project. It builds trust over months. Then they learn it's AI-generated from project management data. The next email feels different. Even if it's just as accurate, just as useful — it's different now.

This is why transparency matters, but it's also why choosing what to automate matters even more. Some things carry weight precisely because a human did them. Automate the thing, and you don't just save time — you remove the weight.

A simple test

Before you automate any process in your business, ask two questions:

1. Would your customer care if they knew?

If a customer found out this task was handled by AI, would they shrug — or would they feel shortchanged? Invoice processing: shrug. Personalised strategy advice: problem.

2. Is this part of why they chose you?

Your customers picked you for a reason. If this process is connected to that reason, it's probably handmade. Protect it.

If the answer to both is no — automate aggressively. Free up every minute you can. That's pure leverage.

If the answer to either is yes — stop and think. Maybe there's a way to use AI as a tool in that process without replacing the human judgement. Maybe AI does the first pass and a person adds the thinking. Maybe you automate 80% of the process and protect the 20% that carries the weight.

But don't just automate it because you can.

Clear intent first

Being AI-ready in 2026 doesn't mean adopting every tool that launches. It means having a clear picture of what your business actually is — what customers value, what they'd miss if it disappeared — and letting that guide every automation decision.

The best yong tao fu stall in your neighbourhood didn't get famous by being efficient. It got famous by being deliberate about what mattered. The uncle knew exactly which parts of the process were his, and he protected them.

Your business has handmade parts too. Know what they are. Protect them. Automate everything else.