The following commentary is contributed by Dataiku Regional Vice President of ASEAN Sales Mochamad Idham
As enterprises look toward 2026, one of the most consequential shifts in artificial intelligence (AI) will not come from new models or faster infrastructure, but from a change in who actually builds and governs enterprise AI. The next phase of AI maturity will be driven by a previously under-recognised workforce inside organisations: Employees who understand the business deeply enough to translate how decisions are truly made into agentic systems.
These individuals are not traditional data scientists. They are domain experts who know the judgment calls, the edge cases and the unwritten rules that never appear in process documents, yet define how organisations really operate. In 2026, they will increasingly take on responsibility for shaping, supervising and refining AI agents that support core business workflows.
This marks the emergence of a new designer role in the enterprise – one that blends business expertise with sufficient technical fluency to work with agentic systems and orchestration layers.
Moving Beyond Pilots to Operational AI
Organisations that recognise and empower this group will be the ones that move beyond proof-of-concept AI projects into scaled, operational deployment. Those that do not will remain stuck experimenting with tools that never reach production. In a market like Malaysia, where the New Industrial Master Plan 2030 and National Fourth Industrial Revolution Policy emphasise technology-driven productivity and responsible digital adoption across sectors, this divide will become increasingly apparent.
When Human and Agent Workforces Collide
Many organisations will enter 2026 without fully acknowledging that they already operate a hybrid workforce. Human employees and AI agents are collaborating inside everyday processes, often without formal oversight or governance. In some cases, agentic systems have been embedded faster than leadership frameworks can keep pace. In fact, according to our latest global report, The 7 Career-Making AI Decisions report for CIOs in 2026, a majority of CIOs (82%) agree that employees are creating AI agents and apps faster than IT can govern them.
This reality is driving a new focus on accountability. Organisations must define who supervises agents, who validates their outputs, and who remains responsible when automated decisions affect customers, financial outcomes or regulatory compliance. Malaysia’s increasing focus on trustworthy AI, demonstrated through frameworks such as the government’s AI Governance and Ethical Guidelines and regulatory standards from bodies like Bank Negara Malaysia underscores why clear accountability in this domain is now essential.
The True Cost of AI Reaches the C-Suite
Another reality becoming impossible to ignore is the real cost of AI. In 2026, pressure from boards and investors will push AI economics firmly onto the C-suite agenda. Our report showed that nearly all CIOs (95%) brief the board on AI performance at least quarterly. We also found that 98% of CIOs report that board pressure to demonstrate measurable AI ROI has increased since 2024. The cost of AI is no longer limited to model licensing or cloud compute. It includes the operational burden of fragmented platforms, the coordination costs arising from overlapping initiatives, and the strategic risk posed by systems that are insufficiently governed or monitored.
As CEOs, CIOs and CFOs demand a clearer picture, many organisations will discover that the largest expense lies in the overhead required to keep AI consistent, efficient, and compliant across the enterprise. Total cost of ownership, not model pricing, will become the primary decision framework for AI investment.
The AI Bubble Isn’t About Value, It’s About Timing
There is growing talk of an AI bubble, but it is not about whether AI is useful. Capital spending on infrastructure is currently outpacing demand, creating a familiar bubble dynamic. The real story is where profits will ultimately land, not with infrastructure-heavy AI providers, but with enterprises that can convert this investment into tangible operational outcomes.
When the cycle corrects, it will be organisations that rapidly translate AI into business value that emerge strongest.
The Hidden Risk of Vendor Lock-In
At the same time, a less visible but more dangerous risk is emerging: Reasoning lock-in. In 2026, the greatest vendor risk will not be cloud or model dependency, but the outsourcing of core business logic to third-party agentic platforms. When enterprises embed their decision-making processes into external systems, they sacrifice adaptability.
Organisations that retain ownership of their orchestration and reasoning layer will remain flexible. Those that do not will face painful and costly migrations as their operating models evolve.
