Singapore – A majority of businesses in Singapore are using or piloting agentic artificial intelligence (AI) systems, but many lack the ability to produce evidence explaining how those systems make decisions, according to a new report released by Sumsub in collaboration with the Singapore Fintech Association (SFA).
The report found that 94% of Singapore businesses are using or testing multi-step AI systems. However, only 29% said they could produce an audit trail for AI-driven decisions, highlighting what the report describes as an “Accountability Asymmetry”—a gap between organisations’ responsibility for AI decisions and their ability to demonstrate how those decisions were made.
According to the report, the lack of traceable AI decision records could expose organisations to operational risks, regulatory scrutiny, and reduced customer trust as AI systems increasingly operate with limited human oversight.
The study evaluated AI governance across Singapore and the wider Asia-Pacific region using three measures: autonomy, which assesses the extent to which AI systems act independently; responsibility, which examines how ownership of AI outcomes is assigned; and traceability, which measures whether AI decisions can be reconstructed and explained.
Among the Singapore findings, 70% of businesses reported having explicit guidelines assigning responsibility for AI outcomes to either an individual (40%) or a team (30%), matching the regional average.
The report also found that 90% of Singapore businesses are comfortable allowing AI to perform low-risk routine tasks, exceeding the APAC average of 88%. However, organisations showed greater caution when AI applications involved financial liabilities.
Only 16% of Singapore businesses significantly expanded the scope or autonomy of their AI systems over the past year, which the report identified as the most measured deployment rate in the region. Singapore recorded an overall AI governance benchmark score of 65.6, slightly below the APAC average of 67.1.
Businesses identified data analytics (29%) and operations or workflow processing (21%) as the areas where AI currently delivers the greatest value, followed by security-related applications such as fraud detection, anti-money laundering, and risk monitoring (15%).
“Everyone is focused on how quickly AI is advancing, but the bigger question is whether governance is keeping pace. Our joint survey with Sumsub found that fewer than one in three organisations can produce an audit trail for AI-driven decisions. As AI moves beyond copilots into autonomous agents handling increasingly critical workflows, the focus now should be on building the traceability, accountability and governance needed to deploy AI at scale. As an industry, that’s where our attention needs to be next,” said Holly Fang, President, Singapore Fintech Association.
The report said Singapore’s governance score reflects the country’s regulatory maturity rather than slower AI adoption. Earlier this year, the Singapore government introduced the Model AI Governance Framework for Agentic AI, allowing organisations to assess their AI governance against technical benchmarks instead of documentation-based compliance alone.
“Prudence, rather than a lack of strategic intent, defines how the enterprises are scaling AI agents,” notes Penny Chai, Vice President, APAC at Sumsub. “When financial liabilities are on the line, immature traceability systems create an unacceptable operational risk. Establishing robust tracking architectures and guardrails is the vital prerequisite to safely deploying high-stakes AI at scale.”
Across the region, Thailand recorded the highest AI governance benchmark score at 70.3, followed by the Philippines at 69.6. India scored 68.5, China 68.0, Hong Kong and Australia both 66.7, Indonesia 66.0, Singapore 65.6, and Malaysia 62.4.
The report also found that financial services led sector rankings with a governance score of 69.6, supported by a 68% audit trail adoption rate. IT and software services followed with a score of 68.8. E-commerce platforms scored 65.4, while the mobility and delivery sector ranked lowest at 64.4.
When asked about the technical barriers to scaling AI, 66% of Singapore businesses cited the complexity of modern AI models as their biggest challenge. Half pointed to system integration issues, while 49% identified the need to develop tracking mechanisms for actions performed by third-party AI tools.
Meanwhile, 98% of respondents said they would be willing to adopt third-party verification solutions capable of linking AI-driven actions to verified identities as part of strengthening AI governance.
“Public-private collaboration is the critical engine for APAC’s AI future. Regulators are laying down the policy blueprints, but the tech sector must step up with the operational plumbing to address system integration and model complexity” Chai concludes. “Long-term success hinges on ecosystems where national frameworks, like MAS’ Safeguards for Agentic Finance at Runtime (SAFR), are powered by industry recognised trust infrastructure. By anchoring automated actions to a secure, human-accountable digital trail, businesses can seamlessly turn compliance into verifiable digital trust.”
