After Early AI Failures, Businesses Ground Efforts in Operations and Governance

After 18 months of rushed AI pilots and forced employee adoption, businesses are discovering that the path to meaningful AI value runs through operational foundations, not flashy experiments. New research from automation vendor ServiceNow, shared at its World Forum event in Sydney, paints a picture of an industry recalibrating — moving from what its Asia-Pacific president Adrian Johnston calls "shallow ROI" in the form of personal productivity copilots toward structured, governed AI agent deployments embedded in core business workflows.

ServiceNow's annual Enterprise AI Maturity Index, surveying 4,500 executives across 19 countries, found AI spending surged 110% year-on-year. Yet the payoff remains elusive for most: only 16% of organisations have replaced fragmented legacy systems with an integrated IT platform capable of supporting AI at scale. For the rare few that have, the results are tangible — automated server issue resolution, role-based access management, and supply chain fixes handled without human intervention.

The Singapore banking sector is a case in point. Johnston noted that banks here may soon need a clear inventory of their AI tools with robust controls throughout the tools' life cycles, as regulatory expectations around AI governance tighten. This mirrors a broader regional trend toward structured AI oversight — ServiceNow's recent acquisition of cybersecurity firm Armis, for instance, targets the growing need for AI agent visibility and vulnerability management in enterprise environments.

The shift carries clear implications for Singapore's financial and tech sectors. Businesses that have leaned heavily on employee-facing copilots are beginning to understand that these tools deliver convenience rather than transformation. The real gains, ServiceNow argues, come from AI agents that execute tasks autonomously within tightly governed frameworks — resetting passwords, managing system access during role changes, or patching server vulnerabilities. But getting there demands integrated platforms, clear governance, and human oversight loops.

Why it matters for Singapore: As a financial hub with an ambitious national AI strategy, Singapore is uniquely positioned to benefit from — and be tested by — this shift from experimentation to production AI. The Monetary Authority of Singapore has already signalled heightened scrutiny of AI deployment in financial services, and the city-state's banks are among the first to face governance requirements around AI tool inventory and lifecycle controls. The message from ServiceNow's research is clear: the winners in AI adoption won't be the ones with the most pilots, but those that build the operational and governance infrastructure to make AI work reliably at scale.