Navigating the CX evolution in Malaysia: Balancing AI efficiency with human empathy

The customer experience is undergoing a massive shift across Malaysia as businesses grapple with the rapid advancement of artificial intelligence. Despite the push to integrate new technologies, the consensus among industry leaders is that human connection and emotional intelligence remain irreplaceable.

At a recent MARKETING-INTERACTIVE roundtable in Malaysia, hosted in partnership with TP and AWS, senior executives from banking, telecommunications, logistics, fast-moving consumer goods and property development gathered to discuss a central challenge: how can brands successfully integrate AI into their CX strategies without losing the human touch?

The current state of AI adoption

While AI dominates industry conversations, the actual implementation reveals a more cautious reality. For many organisations, human agents still handle the vast majority of customer interactions, particularly those involving complex problem-solving or high emotional stakes.

John Lau, senior vice president and head of business optimisation, strategy and solutions, and the contact centre at Alliance Bank, said AI and new technologies present an opportunity for contact centres to rethink how they operate and improve efficiency, operational resilience, and cost management.

“Customers are increasingly willing to adopt self-service, but they expect it to be simple, conversational, and capable of resolving their requests with minimal effort. Success depends on having the right platforms, processes and talent equipped to work with AI, while keeping the experience simple and safe for customers to use,” Lau said.

He noted that as businesses and transaction volumes continue to grow, automation is becoming increasingly important. “Without automation, continuous efforts to reduce friction in our processes and encourage customer adoption of self-service, contact centre volumes, and operating costs will continue to rise.”

Lau added that modern platforms provide insights that help organisations make better decisions to simplify processes, reduce avoidable contacts, and enable agents to work more efficiently which reduces handling times. However, he cautioned that technology alone is not enough.

“If AI is deployed on top of fragmented processes, many conversations may still end up being transferred to agents, creating frustration rather than improving the customer experience, thereby limiting the benefits,” he said.

“The focus should be on getting the foundations right – the platform, data, processes and people. AI works best when supported by strong foundations, good governance and simplified customer journeys.”

Bernard Yong, chief experience officer at Mah Sing Group, shared the distinct challenges faced by property developers. Because development projects are often decentralised with standalone marketing teams, maintaining a consistent service delivery across all touchpoints is difficult.

Yong revealed that Mah Sing Group has rolled out a chatbot that successfully handles about 15% of the total ticket volume, with plans to extend AI capabilities into pre-sales qualifying.

In the telecommunications sector, the adoption is accelerating. Mike Pui, head of sales at Maxis, shared that the company is actively testing and adopting generative AI.

Maxis leverages large language models to assist managers internally and power its customer portals. Pui explained that intelligent AI agents can streamline complex tasks such as planning a multi-country travel itinerary and recommending the best data roaming packages, significantly reducing the effort required from the customer.

Overcoming legacy systems and data silos

A recurring theme throughout the discussion was the critical importance of data hygiene and the hurdle of legacy systems. AI relies entirely on the data it is fed, and fragmented data can derail even the most sophisticated tech implementations.

However, not all automation needs to be a massive, systemic overhaul. Andrew Jeffree, senior partner, business development, Amazon Connect APJ at Amazon Web Services (AWS), suggested that businesses often stumble by trying to build AI for legacy workflows.

“The return on investment model, very clearly, is still a little bit broken ... in terms of implementing AI and seeing the benefits,” he said.

He advised companies to test, trial and fail fast at a cheaper cost rather than buying fully fledged, inflexible products. He also reminded the room that according to Gartner, 70% of customers visit a website before contacting a brand, stressing the need to design better self-serve experiences at the very beginning of the customer journey.

Empowering employees with agent-assist technology

Rather than viewing AI as a replacement for human workers, forward-thinking leaders are leveraging the technology to empower their internal teams.

Romeo Solas, APAC vice president of digital innovation and transformation at TP, detailed how AI can capture the expertise of top-performing agents, share that expertise across the organisation and enable other agents to adopt the same best practices.

“How do you make these heroic conversations repeatable?” he asked. By analysing thousands of actual conversations, companies can identify the intent behind customer queries and provide agents with real-time guidance. Solas explained that agent-assist tools can automatically transcribe conversations, translate languages, direct agents to the right information and provide the “next best actions” based on established policies, drastically reducing the cognitive load on the agent.

This internal focus helps minimise the fear often associated with AI adoption while improving service quality. Andy Rangel, chief executive officer of Thailand and Malaysia at TP, expanded on this ecosystem approach. He noted the same AI technologies used to serve customers can be deployed to support new hires during their most vulnerable early months on the job.

Furthermore, Rangel shared that TP integrates AI across its entire operation, including talent acquisition. By using AI-driven assessments via platforms such as WhatsApp, the company can efficiently screen candidates, drive efficiency, and reduce costs before a human recruiter even steps in.

Localising the future of CX

Looking ahead, the Malaysian market presents unique linguistic and cultural challenges that AI developers must overcome to deliver truly seamless experiences.

Shiao Chan, head of marketing at Spritzer, pointed out the limitations of current AI platforms in handling the complexities of local vernaculars. While text-based chatbots have been the focus for the past five years, Chan believes the future lies in voice AI.

However, the technology must evolve to understand regional accents and mixed languages such as the blending of Malay, English, and local dialects. “The machine needs to learn, and they don’t have enough to learn because no one has developed [it] yet,” said Chan, emphasising the gap in the current market.

Leon Tham, general manager of sales, marketing and customer service at TNTT Packages Express, echoed these sentiments, noting that B2B logistics customers often communicate using local slang and voice notes that current bots struggle to interpret. Despite these challenges, he acknowledged the potential for AI to drive massive efficiencies, especially as clients continuously demand lower operational costs.

Xing Jun Khoo, head of digital, media and insights at Dutch Lady Milk Industries, brought up the critical limitations of AI when handling highly sensitive issues. He provided the example of severe product complaints, where an automated response is entirely insufficient. In situations involving consumer safety, the process demands nuanced human judgment and a strict, empathetic protocol that cannot simply be handed over to an algorithm.

Conversely, the digital banking sector is aggressively pushing the boundaries of what AI can automate. Lam Chern Feei, head of credit at Boost Bank, noted that as a digital-first entity without physical branches, customers already expect AI and digital call centres to be the primary mode of interaction. Boost Bank is currently working on voice AI that supports both English and Malay, with plans to introduce Mandarin and Cantonese.

Terrance Pui, credit underwriting manager at Boost Bank, elaborated on the company’s ambition to fully automate the loan underwriting process using AI. By relying on AI to cross-reference bank statements and evaluate inventory photos submitted by applicants, Boost Bank aims to replace costly physical site visits with instantaneous, automated approvals.

“We want to make a difference, as compared with other traditional banks,” he said, outlining a vision where the entire journey from submission to draw down is seamless and rapid.

Ultimately, the successful brands of tomorrow will be those that use AI not just to deflect queries, but to proactively generate value. By starting small, experimenting internally, and keeping human empathy at the core of a strategy, businesses in Malaysia can navigate the complexities of AI and build resilient and customer-centric operations.

Acknowledgements

This roundtable and article were made possible by our partners TP and AWS.

TP is a global leader in digital business services, helping brands bridge the gap between complex technology and authentic human connection. By unifying advanced AI models with deep emotional intelligence, TP empowers leaders to streamline operations, enhance employee capabilities and deploy tailored, hyper-personalised customer experiences at scale.

AWS provides the agile cloud infrastructure and flexible, usage-based solutions – such as Amazon Connect – that enable businesses to seamlessly integrate these AI models and design end-to-end customer journeys without the burden of fixed costs.

To learn more about how TP can help you increase engagement and elevate your customer experience, click here, or scan the QR code below.