Fifty-seven percent of Southeast Asian businesses remain trapped in what experts call the “frozen middle“—they test or roll out artificial intelligence but cannot expand its use across their operations. The problem does not stem from technology limitations, but from how companies structure their AI strategies, according to research involving 461 senior leaders in the region.
Adobe and Incisiv’s findings show that while many firms launch AI pilots, deploy tools, and create strategic plans, they often struggle when attempting to scale these efforts. The distinction between success and failure hinges on organizational readiness, including defined responsibility, coordinated processes, and seamless integration between business and technology teams.
Among Southeast Asian companies, 40% of AI projects—especially those involving agentic AI—are stalled in this “frozen middle” stage. Transitioning from pilot to full deployment takes six months or longer for six in ten organizations, with nearly a third requiring a year or more. These delays frequently result in lost progress: 48% of companies are still working on their first AI application or have abandoned it after failure.
Even when AI systems are operational, most businesses lack a unified framework to support broader adoption. Without standardized procedures, departments operate in isolation, new pilots launch without alignment, and existing initiatives stall. Only 6.3% of Southeast Asian companies, 27 organizations in total, have successfully scaled AI across multiple functions and built a sustainable model. These top performers convert pilots into production nearly four times faster than their peers.
Investment alone does not guarantee success. While 73% of companies are increasing their AI budgets annually, without proper organizational foundations, these expenditures fail to deliver measurable results. Nearly 60% of executives admit they are still underfunding AI initiatives, yet the core issue remains how deeply these technologies are embedded into daily operations.
Why Leadership Structures Make or Break AI
Leaders in Southeast Asia outperform regional averages in two key areas: structural clarity (dedicated ownership and accountability) and workflow integration (how AI is woven into business processes). Their maturity in these areas surpasses the regional norm by 67% and 59%, respectively. However, most companies still grapple with basic coordination issues. Only 12% have a designated AI owner with budget authority, compared to a global average of 20%. Among the most successful firms, this figure rises to 89%.
Executives reviewing past decisions frequently identify missed opportunities. Many wish they had established clearer performance metrics earlier, secured funding sooner, or appointed an AI owner before deployment. In some cases, these foundational steps were only addressed after pilots had already begun, by which point, the chance to set a progress baseline had passed.
Fifteen percent of AI proposals from Southeast Asian organizations ever reach production. The rest fail at various stages: 36% are rejected during IT evaluation, often due to data or architectural gaps, while 18% are blocked by security reviews. In highly regulated sectors like finance and healthcare, rejection rates nearly double.
Talent shortages persist as a major hurdle, with 81% of businesses reporting no improvement, or even a decline, in their workforce capabilities over the past year. Over half are expanding existing roles to include AI responsibilities, but 21% are introducing entirely new skill requirements. Without clear paths to implementation, even promising ideas struggle to advance.
Global leaders extract 1.4 to 1.6 times more value from the same AI technology than their peers, the study notes. The advantage lies not in the tools themselves, but in how they are integrated into operations. Without clear ownership, measurable outcomes, and cross-functional collaboration, companies risk wasting investments, and falling further behind competitors.
How Regional Leaders Outpace Peers
When comparing Southeast Asian firms to global peers, regional executives report higher-than-average confidence in their AI strategies, yet their execution often lags.
Regional leaders demonstrate stronger foundations in two critical areas: data readiness and cross-functional collaboration. Sixty-one percent of top performers have integrated AI into their data governance policies, ensuring consistency and compliance. In contrast, only 28% of less mature organizations have taken similar steps.
Industry differences further illustrate the challenges of scaling AI. In financial services, where compliance and risk management are essential, 42% of proposals are rejected during security reviews, double the rate of other sectors. Healthcare organizations face comparable obstacles, with 38% of AI initiatives stalled due to regulatory or ethical concerns. Even in less regulated industries, 29% of proposals fail because business teams cannot demonstrate a clear link between AI tools and measurable business outcomes.
Government Policies Accelerate, or Stall, AI Growth
Singapore and Malaysia lead the region, with 35% and 33% of companies, respectively, having scaled AI across multiple workflows. These markets share a common factor: government-led digital transformation initiatives that have accelerated private-sector adoption by providing standardized frameworks for data sharing and security. In contrast, Indonesia and the Philippines have lower scaling rates (18% and 16%, respectively), where fragmented regulations and slower digital infrastructure create additional barriers.
Organizational Factors Outweigh Technology in AI Scaling Success
A study of 461 senior executives in Southeast Asia found that unclear ownership, poor coordination, and weak workflow integration are the primary obstacles to expanding AI use, rather than technical limitations. The research, conducted by Adobe and Incisiv, determined that organizational readiness is a stronger indicator of successful AI adoption than the technology itself. While companies may invest equally in models, teams, and budgets, their outcomes vary based on how they structure decision-making, measure results, and align business and IT teams.
Approximately 79% of companies in the region are classified as nascent, meaning they are still developing both AI deployment and operational maturity, higher than any other Asia Pacific market. The findings coincide with a trend where 73% of firms increase AI spending annually, yet nearly 60% acknowledge underinvestment in foundational processes. This gap highlights the need for structured frameworks to convert investment into scalable, measurable value.
Executives identified key missed opportunities, including delayed budget approvals, late assignment of AI ownership, and insufficient change management support. Many addressed these issues only after pilots began, losing the chance to establish early performance benchmarks. Without standardized integration methods, security protocols, or dedicated budget categories, proposals frequently stall, 36% fail IT evaluation, and 18% are rejected in security reviews, doubling in regulated industries like finance and healthcare.
