Daythree AI Labs CEO: AI Won’t Take Your Job, Someone Who Uses AI Will
Daythree AI Labs CEO Chiew Sin Kwang breaks down why Malaysian consumers get frustrated with business AI, whether AI is really coming for people’s jobs, and what SMEs should know before adopting it.
- AI won't replace workers outright; those who learn to use AI tools will replace those who don't.
- Consumer frustration with business AI stems from poor service design, not the technology itself.
- Businesses should start by automating repetitive, low-value tasks with manageable one-off setup costs.
Subscribe to our FREE Newsletter, or Telegram and WhatsApp channels for the latest stories and updates.
Chiew Sin Kwang, chief executive of Daythree AI Labs – a local B2B AI company that builds its systems in-house for clients – says most consumer frustration with business AI comes down to service design, not the technology itself.
The most common consumer complaint, he said, is being stuck with rigid, multiple-choice chatbot menus instead of open-ended conversation.
He broke down AI-customer interactions into three categories: queries AI can fully resolve on the spot, queries where AI can only point customers to next steps, and complaints AI cannot resolve at all.
Problems arise, he said, when companies push customers into automation for that third category, especially after a service has already failed once.
“That’s where the frustration begins. You already fail me, don’t go put me into an automation, I want to speak to him,” Chiew said, describing the customer’s likely reaction.
Language is another point of emphasis. Chiew said many companies feed AI systems old documents written for human readers rather than rewriting them for AI, causing outdated or conflicting information to surface as customer-facing answers.
He added that Malay-English “Manglish” phrasing, and terms like “tapau,” can still confuse multilingual models today, even ones that technically support dozens of languages.
On some of the worries that people might have about AI, Chiew rejected the notion that humans will be replaced outright.
“AI basically will not replace the person. That person will be replaced by someone who knows how to use AI,” he said.
He cited his own team as an example, describing a coding challenge he ran against one of his developers – with Chiew using Claude, and the developer coding manually.
Chiew said he was faster than the developer despite not knowing how to code manually, because AI levelled the playing field between someone directing the work and someone executing it line by line.
He said workers who learn to use AI tools can take on more projects and become more valuable, while those who don’t risk falling behind.
Businesses Need To Take the First Step on AI
For businesses unsure where to start, Chiew recommended automating repetitive, low-value tasks first, such as sorting emails or sending payment reminders, rather than attempting a costly overhaul immediately.
He said setup costs for simple automations can be a one-off expense of around RM2,000 to RM8,000, compared to hiring and retaining staff for the same repetitive work over a year.
Daythree uses a Six Sigma-based methodology with clients to measure current workloads, define the AI implementation needed, and calculate returns before rollout, according to Chiew.
He said hidden costs are often overlooked, particularly ongoing AI token usage, which businesses should budget for based on realistic customer volumes and seasonal spikes.
Chiew also cautioned against choosing AI vendors purely on price. “Cheap doesn’t mean it’s secure,” he said, warning that low-cost providers may compromise on data protection.
On Malaysia’s national AI push, Chiew said the government could do more to attract global AI infrastructure partners to set up locally, which he said would help both the economy and long-term pricing stability for AI token costs.
In 2025, Daythree AI Labs partnered with Vietnam-based FPT Software, which took a stake in the company as part of a broader push into the Asia Pacific market.
READ MORE: LinkedIn’s AI Problem: Study Finds Majority Of Posts May Be Machine-Written
Share your thoughts with us via TRP’s Facebook, X, Instagram, or Threads.
