Only 15% of organisations in Southeast Asia describe themselves as digitally mature, even as nine in ten employers say they are actively building AI-ready cultures, according to new research commissioned by Lark.

The report, titled “The Paradox of Progress — Why a Broken Employee Experience is Sabotaging Adoption of AI in the Workplace”, is based on a double-blind survey of 900 employers and more than 5,000 employees across Singapore, Malaysia, Indonesia, Thailand, the Philippines and Vietnam, conducted by an independent third party. It finds a widening gap between leadership’s AI ambitions and the fragmented daily reality facing employees — a gap the report calls a “Critical Inflection Point”.

Four fault lines in workplace AI adoption

The research identifies four recurring problems. Investment skews toward IT (69%) and Finance (60%) departments, while Employee Experience (48%) and HR (47%) lag behind. More tools have not meant more productivity: 55% of employees lose three or more hours a week to collaboration inefficiencies, and 71% feel overwhelmed by the number of platforms they must check.

A gap between stated intent and lived autonomy also stands out. While 80% of employers say they support empowerment, only 28% of employees feel they have real autonomy to introduce new ideas. Meanwhile, employers are 17 percentage points more likely than employees to feel comfortable with AI-enabled tools, pointing to an uneven training gap across the workforce.

“These findings should be a wake-up call. We are at a pivotal moment for AI adoption across Southeast Asia, but what this research tells us is that the foundation isn’t as solid as leadership believes,” said Olivier Adam, General Manager, Asia Pacific at Lark.

Trust gap complicates the picture

Transparency is also a sticking point: just 22% of employees believe their organisation is being very transparent about how AI and automation are being deployed, and 55% say leadership’s expectations for AI use remain unclear. 62% worry AI could eventually make their roles obsolete, while 76% hold security reservations about wider AI use.

Yet the data suggests employees are not opposed to AI itself — 88% say they want it to take over routine tasks, and only a slim majority (52%) still trust human-generated output over AI. The report frames this as a demand for clarity rather than resistance: employees want defined boundaries and honest communication about what AI means for their roles.

Organisations that have consolidated fragmented tools onto a single platform report measurable gains, according to the study, with 89% seeing an immediate efficiency boost and 87% reporting reduced communication friction.