Multi-agent adoption to rise 58% by 2027
MuleSoft said governance and integration remain key concerns as more AI agents are deployed across enterprises.
Multi-agent adoption is expected to rise 58% by 2027 as organisations move toward greater use of AI agents, according to MuleSoft’s report.
Singapore organisations currently use an average of 12 AI agents, but half of these agents operate in isolated silos rather than as part of a multi-agent system.
MuleSoft said this could lead to disconnected workflows, redundant automations, fragmented infrastructure, and shadow AI.
The report found that 96% of Singapore IT leaders believe AI agent success depends on seamless data integration across all systems. Meanwhile, 97% said AI agent success will require IT architecture to become more API-driven.
Singapore enterprises use an average of 1,002 applications, but only 27% are integrated. As a result, 92% of IT leaders are concerned that agents may introduce more complexity than value.
The top challenges to agentic transformation are integrating siloed apps and data, cited by 45% of respondents; resources and budget allocation, cited by 44%; and legacy infrastructure or system incompatibility, cited by 43%.
Data readiness is also a concern, with 98% of organisations facing barriers to using data for AI use cases. Around 41% identified outdated IT architecture or disconnected systems as a top blocker.
Governance remains a major issue, with 62% citing cross-application data governance as a top integration challenge. An estimated 31% of APIs are currently ungoverned, whilst only half of organisations have a centralised governance framework for agentic capabilities.
Still, Singapore IT leaders expect AI agents to deliver benefits. About 98% said agents have improved or are expected to improve employee experiences, whilst 95% believe they will free developers to focus on higher-value work.
The Singapore findings are based on responses from 100 IT leaders as part of Salesforce’s 11th annual Connectivity Benchmark Report.