Enterprise artificial intelligence (AI) is leaving the sandbox, shifting from back-end pilots to large-scale, customer-facing deployments.
Global consultancy Accenture said success hinges on three pillars: robust architecture, governed knowledge and workforce transformation.
"2026 will mark a turning point where AI adoption moves from experimentation to large-scale deployment, as enterprises begin launching more customer-facing AI solutions and agentic AI systems into real operations," Anoop Sagoo, chief executive of Southeast Asia at Accenture, told the Bangkok Post.
The AI industry has evolved rapidly to focus on how companies can scale AI adoption and generate measurable business value.
THREE BARRIERS
However, three major barriers are still slowing enterprise-wide implementation. The first challenge is building a strong AI foundation, especially in data infrastructure, cloud migration and application modernisation, he said.
While AI pilots are easy to run in isolation, scaling AI in live operations requires integrated cloud systems, standardised data environments and modern applications that can use data effectively.
The second challenge is creating enterprise knowledge bases or "AI brains". AI systems need deep organisational context, including procedures, workflows and policies, to operate accurately in business settings.
For example, conversational AI must align with internal compliance rules and customer service processes, yet many firms still lack the structured knowledge systems required.
Mr Sagoo said the third obstacle is governance and workforce transformation. AI adoption is not just about automation; it requires redesigning workflows, retraining employees and putting guardrails in place for responsible and secure use.
"Many companies still underestimate the scale of change management needed," he said.
Agentic AI marks a major evolution beyond traditional AI. Agentic AI autonomously takes action and coordinates multiple agents to execute complex tasks -- from running marketing campaigns to optimising supply chains.
The technology is becoming central to enterprise transformation, enabling organisations to completely redesign operations and unlock entirely new capabilities.
Sovereign AI is surging across Southeast Asia as governments race to control their own data, models and infrastructure amid geopolitical friction and data residency fears, said Mr Sagoo.
While Singapore leads in terms of administrative oversight and advanced policies, momentum is shifting regionally, he noted. However, Malaysia is questioning if resource-heavy foreign data centres offer enough economic payback, while Indonesia remains focused on keeping data within its borders.
Thailand is leveraging its strategic position to attract both Western and Chinese AI giants.
As AI tools advance rapidly, companies making bold, scaled investments are pulling ahead. Those that have scaled at least one strategic AI initiative are nearly three times more likely than peers to see AI returns exceed expectations, according to Accenture.
Mr Sagoo urged companies to focus on business value over disconnected AI pilots, prioritising high-impact use cases with clear outcomes, especially customer-facing applications that drive growth and competitiveness.
He said AI adoption also requires rethinking operating models, as AI-driven workflows blur traditional functional boundaries and connect the full customer journey.
Mr Sagoo recommended a "build and buy" approach, combining external AI platforms with selectively developed proprietary capabilities for strategic functions.
"The future AI landscape is expected to involve hybrid environments combining multiple AI platforms, cloud providers and internally developed models," he said.
Accenture is driving its own internal AI transformation, deploying 70 to 100 AI agents across human resources, finance and marketing.
In Thailand, the firm recently used AI to screen 7,000 internship applications for just 70 roles.
SE Asia AI race
The company ranks Singapore as Southeast Asia's most mature AI economy, with Thailand second due to its early national AI strategy and strong enterprise adoption.
Malaysia and Indonesia are moving fast, but remain in earlier execution phases.
"Thailand should focus more heavily on AI workforce skills and real-world implementation to build long-term competitiveness," said Mr Sagoo.
Banking is the clear leader in AI adoption, driven by heavy technology investment and rising competition from digital and virtual banks, though telecom, retail and energy are high-potential sectors, he noted.
Thailand's strong consumer market could benefit from lessons drawn from China's AI-enabled retail innovations, said Mr Sagoo.
In energy, AI can improve asset management and predictive maintenance by analysing video, operational and sensor data to predict equipment failures and optimise maintenance schedules, unlocking efficiency gains and attracting further technology investment.
"Thailand has strong ambition to use AI to improve the customer experience, strengthen operational resilience and unlock new growth, but success will depend on connecting strategy with execution and building the foundations needed to scale impact," said Patama Chantaruck, Thailand managing director at Accenture.
