Talk, don’t tap: How ASEAN customers will engage brands

One of Salesforce’s 2026 predictions for ASEAN states that there will be a wave of local and region specific AI and small language models tailored specifically to ASEAN’s diverse languages. This is significant as this region’s language diversity is often seen as a hurdle to digitalisation at scale. 

According to Gavin Barfield, Salesforce’s VP and CTO of solutions in ASEAN, “The first use cases will be high-volume workflows, because that’s where language nuance has an outsized impact on resolution, trust, and adoption. This is also why we’re making Agentforce available in Southeast Asian languages such as Thai, Vietnamese, Bahasa Melayu and Bahasa Indonesia, so businesses can deploy AI agents that work naturally in-market.”

Customer service is the obvious starting point, but he expects many organisations to begin on the inside. “Many organisations will choose to start with employee-facing workflows as a lower-risk way to establish the right governance and operating discipline before extending agentic experiences to customer-facing, high-volume use cases,” he said.

Leapfrog

Yet another prediction, Salesforce made, revealed the organisation’s ambitions.

Gavin argued that this region can skip legacy IT and go straight to agentic AI workflows.

 “Every business stands to benefit from leveraging agentic AI,” Barfield says, but he singles out micro, small and medium enterprises (MSMEs) and less encumbered organisations as the real early winners.

“There is a significant opportunity for MSMEs. MSMEs traditionally face resource and capacity constraints that slow operations and limit customer engagement. AI agents can offer a scalable workforce, allowing small businesses to match the service capabilities of larger firms and reduce inefficiencies,” he explained.

Those dynamics are already visible in service-heavy industries. “Organisations with large volumes of customer interactions and repeatable workflows will be best positioned to leverage agentic AI. Early adopters are financial services, travel and aviation, telecommunications amongst other service-led industries where speed and consistency directly translate to customer satisfaction and business growth,” he said.

Barfield points to concrete deployments to show that this is more than theory. “In ASEAN, Singapore Airlines (SIA) is incorporating Agentforce and Data360 into its customer case management system, enabling it to deliver more consistent and personalised service to its customers,” he noted. 

“Additionally, Singlife is adopting Agentforce to provide its customer service executives with access to an intelligent AI agent capable of delivering real-time, trusted responses to a broad range of product-related questions.” Globally, references like OpenTable and Engine provide further proof that AI agents can lift productivity and consistency at scale.

Voice, channels, and the new interface

According to Gavin, with advances in voice technology, AI agents will increasingly be able to handle high-volume customer calls in real time, reducing reliance on Interactive Voice Response (IVR) menus (e.g., “Press 1 for sales”) and hold music – workarounds that exist because service capacity has historically been constrained by manpower.

In ASEAN, Gavin expects voice-first agentic workflows to take off fastest in customer service; agents can act as a conversational triage like understanding intent, routing customers to the right team, and increasingly resolving routine requests.

At the same time, he opined that channel strategy itself needs to be rethought. “Channel strategy should shift from ‘help customers navigate apps and menus’ to ‘let customers express intent conversationally in whichever channel they choose,’ with consistent outcomes across voice, chat, messaging and web.”

With that shift, escalation design and guardrails become central. “IT and CX leaders need to prioritise clear escalation rules and seamless handoffs where human agents inherit full context, plus enterprise guardrails – permissions, transparency and auditability – so leaders can trust how the agent behaves as it scales,” he added.

Proofs-of-concept: 2024 till today

In 2026, success for ASEAN businesses will hinge on successfully moving beyond pilots to operational deployment: agents used in day-to-day customer and employee workflows, in local languages, with clear ownership, measurable outcomes, and consistent quality across channels.

The difference from many 2024–2025 pilots is that early efforts often proved that LLMs could work, but only in narrow scenarios. This year, success will be defined by closing the last mile to mass adoption. This means grounding agents in trusted enterprise context and data, applying deterministic controls and permissions, and running with observability and orchestration so outcomes stay consistent across markets and channels.

With this foundation in place, local language support then becomes an accelerator, allowing teams to reap the benefits of local language support immediately.