Much of 2025 has seen the word “digital employees” enter our vocabulary, especially as the noise surrounding agentic AI has become louder. Some consider them replacements, while many providers consider them augmentation agents.
“We view AI agents as ‘digital employees’ that extend the capabilities of human workers by autonomously executing defined business processes,” defines Jess O’ReillyDirector General of ASEAN at working day.
But the problem is not the definition. Rather, it is a question of whether treating digital employees means treating them equally to human employees. And this difference creates a paradox in Singapore’s corporate towers – one that should concern every executive planning their AI strategy.
So while 82% of organizations are rapidly expanding their use of AI agents in a recent Workday study, employees have drawn a remarkably clear red line in the digital sand: 75% are comfortable working alongside AI agents, but only 30% feel comfortable being managed by them.
The headache of agent sprawl
The craze for AI agents creates a new operational headache that O’Reilly calls “agent sprawl.” This is the case when organizations deploy large numbers of AI agents without clear roles or oversight, creating security, compliance and cost management risks. It’s the problem with SaaS expansion all over again, except this time the tools can make autonomous decisions.
In Singapore in particular, the barriers are particularly complex. According to Workday’s study, 41% of respondents see ethics and governance issues as the top barrier to adoption of AI agents, while 33% cite compliance with changing regulations and data security and privacy vulnerabilities as top concerns in finance functions. These figures reflect Singapore’s strict regulatory environment and the city-state’s position as a regional financial hub where data governance is not optional.
“Organizations in Singapore need a centralized way to manage and govern their AI agents, so they can harness their potential while maintaining control,” says O’Reilly. This is where the emerging concept of an “agent of record system” becomes critical, which is essentially an HR system for your digital workforce.
The trust gradient
The data also highlights something interesting: trust in AI agents is not binary but contextual and very specific. The Workday study reveals a clear gradient in trust: users were more comfortable with AI agents recommending skill development, but less comfortable with AI agents making or directly managing financial decisions.
The pattern is revealing. Workers trust AI to power IT infrastructure and technology, where speed and accuracy of data are most important. They are more comfortable with AI agents that recommend skill development. Yet when decisions involve people, money, or respect for the law, trust shifts decisively toward humans; only 41% are comfortable with AI making financial decisions.
“AI is never intended to replace human decisions, but rather to augment them,” O’Reilly emphasizes, describing Workday’s “human-in-the-loop approach,” which ensures human oversight at key decision points. A strong majority in Workday’s latest Agentic AI study also supports regulating AI agent autonomy through ethical guidelines set by developers (57%) and strict human oversight (48%).
Operational transformation
The implications for organizational structure are profound. When you introduce AI agents into an HR environment, O’Reilly notes, “it creates an opportunity for greater collaboration across departments, placing greater responsibility on HR teams to lead the conversation around workforce tools.”
This creates a fundamental shift in decision-making authority. Traditionally, the operations manager and the IT manager shared decisions about staff tooling. Now, HR must take ownership of the data and processes that AI agents rely on, establishing what O’Reilly calls “a single source of truth through a centralized repository of employee data.”
Think of it as the emergence of the “head of work,” a role that transcends traditional HR boundaries to encompass the design of how work gets done. It optimizes the performance of human-AI teams and provides support to employees as AI becomes an integral part of the workforce.
The Singapore context
For Singapore-based organizations, the challenge is compounded by workforce skills readiness. According to a Workday study, only 30% of leaders in Singapore are confident in their workforce’s ability to acquire skills. AI-powered skills intelligence tools can help close this gap, but only if organizations invest in the required change management.
“Beyond implementing technology, leaders need to help employees unlearn old practices and adopt new systems,” observes O’Reilly. “Cultivating a positive culture is essential to supporting employees through this transition. »
Rather than simply training in new tools, this culture must be about fundamental organizational change. For example, the report notes that in finance functions, 72% believe AI agents could help address the CPA shortage and fund professionals, but only if their organization builds and maintains trust.
The new future of work
The most successful implementations will be those that respect the boundaries set by employees. Use AI agents for what they are good at: data analysis, process automation, pattern recognition. But save human judgment for decisions involving empathy, nuanced understanding, and ethical deliberation. A large majority of business leaders and employees (89%) surveyed in Workday’s latest Agentic AI study agree that AI agents will increase productivity.
“AI agents will serve as collaborative tools that support and enhance human decision-making, particularly in sensitive or high-stakes business functions,” says O’Reilly. This means designing systems where AI recommendations can be ignored, where regular independent audits detect bias, and where employee feedback loops quickly detect unexpected results.
JLL, for example, deployed HiredScore AI for recruitment to improve its hiring processes. It reduced average time to hire from 52 days to 48 days and increased quarterly hiring by 64% after implementing AI for recruiting, O’Reilly says. But they also repositioned its talent acquisition team from a largely reactive function to a more strategic role, rather than simply cutting positions.
The future of work rests less on pitting humans against machines than on establishing a good division of labor between them. Organizations that achieve this balance will gain a competitive advantage. Those who fail to do so will find themselves faced not only with agent proliferation, but also employee resistance, erosion of trust, and ultimately failure of AI investments.
The choice, as always, is up to the organization. But employees have already made their position clear: they will work with AI agents, but only on their terms.
Image credit: iStockphoto/Panya7
