Deloitte report finds AI progress uneven but accelerating

Deloitte's 2026 enterprise AI survey shows companies are rapidly experimenting with AI, but turning pilots into scaled production systems remains a challenge.

Deloitte's AI Institute has released its 2026 State of AI in the Enterprise report, offering a snapshot of how organizations are using artificial intelligence and what challenges come with it.

The study draws on responses from more than 3,000 senior leaders involved in AI programs, including 75 in Singapore. It looks at how companies are moving beyond early experimentation and what leaders may need to address as AI becomes part of day-to-day operations.

Chris Lewin, AI & Data Capability Leader at Deloitte Asia Pacific, said: "Business leaders in Southeast Asia have strong AI ambitions, and they are already reaping benefits, particularly in terms of productivity gains. As they continue to invest in AI, they will need a clear roadmap to guide their transformation beyond incremental optimization, with effective governance firmly embedded at all levels. The full value of the technology will come from reimagining what is possible to achieve strategic differentiation and enduring competitive advantage."

Moving past experimentation

One of the report's main themes is the challenge of turning AI pilots into working systems. While many companies are testing AI, fewer have fully rolled it out. In Singapore, 32% of surveyed leaders say at least 40% of their AI pilots have reached production, compared with a global average of 25%. Over the next six months, 54% of respondents — both locally and worldwide — expect to hit that level.

Organizations face a balancing act. They must keep existing systems running while funding AI work that could shape future operations. The report warns that repeated small pilots without scale can slow progress. A clear plan for expanding successful projects may help reduce this "pilot fatigue."

Value beyond efficiency

Singapore leaders report strong early returns. About 73% say AI has improved efficiency and productivity, higher than the global figure of 66%. Just over half say AI has strengthened decision-making by providing clearer data insights.

Even so, large-scale transformation remains limited. Only one-third of Singapore respondents say they are redesigning major processes around AI while keeping their business model intact. Fewer still use AI to fully rethink how their core operations work.

Regulation is the most cited barrier to deeper adoption in Singapore, followed by skills gaps, high costs, and infrastructure limits. To prepare employees, companies are focusing on improving general AI understanding. Many are also reworking career paths as roles shift.

Governance becomes critical

Interest in agentic AI — systems that can take action with limited human input — is rising. Around 72% of Singapore companies expect to deploy it across several operational areas within two years. Leaders see early use in customer service, logistics, and marketing.

Yet governance appears less mature. Only 14% of Singapore leaders report having strong oversight models for agentic systems. Because these tools act rather than simply advise, the report says organizations may need tighter guardrails, real-time monitoring, and detailed audit trails.

Physical AI, which connects software intelligence to machines and control systems, is also gaining traction. Most Singapore businesses expect to use it within two years. Leaders point to digital twins, collaborative robotics, and intelligent monitoring as areas with potential impact. Trust will depend on reliability, system security, and the ability to handle disruptions.

Data location and infrastructure control are becoming strategic concerns. Over three-quarters of Singapore respondents say in-country data handling is important to their planning. Many also worry about reliance on foreign AI infrastructure.

The report suggests companies review which data must stay local, how systems meet regional rules, and how cross-border flows are managed. Clear policies and supporting infrastructure will likely play a central role as regulations tighten.