KUALA LUMPUR (May 8): Malaysian companies that have not begun deploying agentic artificial intelligence (AI) within their workflows risk being left behind, executives from Databricks and Microsoft said at an industry event here.
Speaking at Databricks AI Day 2026 at the Hilton Kuala Lumpur on May 5, Databricks assistant vice-president for field engineering Kunal Taneja said the technology had moved past the experimentation stage and was now ready for enterprise-wide deployment.
“I do think the moment for agentic AI is here. It’s unavoidable, and I think every one of you probably should already be thinking about how you will incorporate agentic AI and moving on that journey in this year as well,” Taneja told delegates.
Taneja said recent studies pointed to productivity gains of at least 30% to 40% from organisations using agentic tools, particularly in surfacing insights from data and in software engineering workflows.
He urged companies to modernise their data architectures to support the shift, and flagged data governance as the single most important enabler. Governance, he said, could no longer be treated as an afterthought confined to row-level or column-level access controls.
“Hopefully it’s become very apparent that governance is actually what’s going to drive most of your AI journey. It will give it that context. Without this context, you cannot make these decisions,” he said.
Taneja also told the audience to begin planning for operational analytics and agentic memory infrastructure, noting that organisations would soon be running large numbers of agents in parallel.
Microsoft country leader for cloud solutions Vivek Chatrath, who shared the stage, said adoption of human-to-agent interaction was already widespread among large enterprises globally, and that Malaysian firms still on the sidelines were behind the curve.
Citing IDC data published in November 2025, Chatrath said an average of 36% of companies across industries were already using agent-to-human interaction at the time the data was collected, with another 50% were in the planning phase.
The IDC data also stated that 88.1% of organisations across financial services are currently using or in the planning stage to deploy agentic AI.
Because the underlying data was gathered roughly nine months before publication, he said the actual share of adopters today was likely significantly higher.
“If your organisation is still thinking about, should I get started, should I not get started, the time is now,” Chatrath said.
He, however, cautioned that agentic AI projects frequently failed when companies put AI engineers to work without first getting domain experts, compliance and risk personnel, and technology teams around the same table.
“Without these three personas, your projects around agentic AI will bomb,” he said, adding that data security lapses involving sensitive enterprise information could stall an organisation’s agentic roadmap by five to six months.
The half-day event drew Databricks customers including PETRONAS and Malaysia Aviation Group (MAG), who took the stage to showcase how they were applying AI and analytics tools built on the Databricks platform across their operations.