Agentic automation enters the operational mainstream in Thai banking

At a recent roundtable hosted by IBM in Thailand, banking and technology leaders discussed how agentic automation is moving beyond process efficiency towards a more strategic role in resilience, customer delivery and operational control.

 Agentic automation is emerging as the next step in how banks think about operational capability, moving beyond scripted process support towards more adaptive systems embedded across core infrastructure.

At the IBM session in Bangkok, speakers from the banking and technology sectors examined how automation is being pushed closer to the centre of financial services, shaped by customer expectations of immediacy and the growing complexity of modern technology estates.

Thailand’s banking sector has become an active testbed for digital delivery, supported by real-time payments infrastructure and a competitive retail market in which mobile-first service is now standard. That environment is raising expectations around uptime, speed and operational resilience.

According to Krungsri Research Institute, Thailand recorded more than 147 million internet and mobile banking accounts by the end of 2024, generating 36.5 billion transactions worth THB 112.8 trillion (around USD 3.1 trillion) over the year. In that context, automation is increasingly being positioned as core infrastructure rather than incremental efficiency.

Participants included Pattarakit Saingarm, assistant vice president for enterprise data and AI strategy at Kasikornbank; Dr. Sanphet Sukhapesna, head of integrated risk management and analytics at Kasikornbank; Anothai Wettayakorn, managing director and technology leader at IBM Thailand; Aju Murjani, automation leader for ASEAN at IBM; and Jerome Walter, field CTO for Asia-Pacific and Japan at HashiCorp, an IBM company.

Together, they explored how banks are moving through distinct stages of automation maturity — from basic process automation to more intelligent orchestration across systems and channels.

Murjani framed AI adoption as outcomes-driven rather than technology-led. “AI for us… is all about what business outcomes you can drive with technology investments you have in your bank,” he said.

Resilience beyond recovery

A consistent theme was that resilience in banking is no longer confined to disaster recovery planning. Institutions are being pushed to design systems that remain dependable through disruption, rather than relying solely on restoration after failure.

Walter argued that resilience must be treated as an adaptive capability, not simply a recovery plan. “Resiliency is not only about how do you bring your service back up when it’s down — it’s your ability to adapt… and to resume services faster,” he said.

That shift matters in Thailand’s competitive digital market. According to payments data providers, systems such as PromptPay collectively process more than 75 million transactions per day, illustrating how central real-time payments and always-on services are to everyday banking.

With large Thai banks such as Kasikornbank managing complex technology estates alongside continuous digital demand, operational resilience is increasingly being treated as a design requirement rather than a recovery function.

Discussion centred on practices such as immutable infrastructure, automated certificate rotation and intentional failure testing — approaches intended to ensure teams respond predictably under stress rather than improvising during outages.

Certificate management deadlines were cited as one example of how operational risk is increasingly shaped by infrastructure discipline as much as by traditional cyber controls. Walter noted that from March 2026, browsers will begin rejecting certificates older than one year, tightening operational expectations around automation and renewal cycles.

Hybrid cloud design was also positioned as central to this resilience agenda, particularly for institutions managing legacy estates alongside rising expectations for always-on digital service delivery.

Customer expectations and the no-wait economy

The roundtable also highlighted how customer behaviour is compressing the cost of downtime. As onboarding journeys become faster and more seamless, tolerance for disruption has narrowed. Service outages are no longer treated as occasional inconveniences, but as moments when customers reassess provider loyalty.

“Nobody wants to wait, so everything has to be quick,” Murjani said. “If your system goes down, we change our behaviours.”

That pressure is reinforced by broader digital adoption: as of the end of 2024, around 90% of Thailand’s adult population owned a smartphone, and internet usage remains high, illustrating how deeply digital services are embedded in daily life, as reported by Krungsri Research Institute.

Case studies referenced included Trust Bank’s rapid digital onboarding model, illustrating how speed and convenience are resetting expectations for retail banking services. In such an environment, disruption becomes a competitive vulnerability rather than a technical inconvenience.

In Thailand, where customers are accustomed to instant payments and app-based banking as a default, service disruption can carry an outsized reputational cost.

Participants noted that AI-enabled automation is increasingly tied to customer satisfaction measures, with banks assessing not only cost reduction but service responsiveness, retention and trust.

From fragmented tools to unified platforms

Operational complexity remains a major constraint. Speakers observed that many institutions still operate fragmented integration environments, with separate tooling for APIs, events, workflows and security layers accumulated over time.

Murjani argued that integration strategies must account for realities beyond API-first thinking. “We realised it wasn’t APIs alone,” he said. “It was APIs, it was events, it was B2B.

Rather than treating integration as a series of disconnected projects, participants discussed the need for more unified operational oversight — particularly as banks expand connectivity across payment networks, merchant ecosystems and third-party platforms.

HashiCorp’s infrastructure tools — including Terraform and Vault — were referenced as examples of the underlying systems banks are increasingly relying on to standardise provisioning, manage secrets and reduce configuration drift across hybrid environments.

For Thai institutions operating at scale, the challenge is less about adopting new tooling than reducing operational fragmentation: fewer handoffs between systems, clearer visibility across dependencies, and tighter control over failure points as digital delivery becomes more complex.

Governance and the move towards agentic systems

As automation becomes more autonomous, governance frameworks were framed as necessary enablers rather than barriers.

Participants emphasised the need for explainability, ethical controls and human oversight as AI moves closer to decisioning in regulated financial environments. Walter stressed that autonomy does not remove the human role. “We may be entering an age of AI — but the human remains the most important in times of crisis,” he said.

Longer-term security considerations, including the eventual implications of quantum computing for encryption, were also noted as emerging risks that institutions will need to prepare for over time. The session positioned agentic automation as the next stage in banking’s operational evolution — where systems can learn, adapt and support predictive risk identification, rather than simply executing repetitive tasks.

Examples such as BRI’s reported onboarding of 85,000 urban youths were cited as illustrations of how AI-enabled automation can expand reach and support measurable business outcomes.

A practical shift in automation’s role

The discussion in Thailand reflected a broader operational shift: automation is moving away from isolated productivity gains and towards infrastructure-level importance.

For banks, the challenge is not whether AI will be adopted, but whether it can be deployed in a way that strengthens resilience, supports governance and improves customer delivery without adding fragility. For Thai financial institutions operating in one of Asia’s more digitally mature retail markets, the operational test is becoming sharper: automation must strengthen resilience and control, not introduce new points of fragility.

The case put forward was that agentic automation is moving into the infrastructure layer of banking — where it shapes uptime, integration discipline and operational decision-making, not just efficiency metrics. The next stage will be defined by whether banks can govern autonomy with the same rigour they apply to risk and resilience, as customer tolerance for disruption continues to narrow.