New Research Reveals the Reason Work Feels Harder Than It Should

In this article we discuss:

• Employees are engaged, but buried in manual integration
• AI is more effective within workflows than as a peripheral tool
• AI adoption increases when the system is trustworthy

A new report from Workday finds employees across Hong Kong and Taiwan are positive about work and AI, but when it comes to tasks that cross systems or departments, things fall apart.

It turns out many people are wasting time acting as the glue between disparate systems, limiting the impact of AI in the enterprise.

AI was expected to free people to do more strategic and creative work, but around 8 in 10 employees in the region say they spend significant time coordinating across teams, moving information between tools and reconciling conflicting data.

The research highlighted that:

• 90% of employees say AI has improved their day-to-day work experience
• 69% spend at least half their time translating and coordinating between systems and teams
• Only 21% of organizations have embedded AI into the core of their business

By adopting AI tools that do not have direct access to core business data and processes, companies have buried their employees in busywork: translating, copying and pasting, double-checking, and coordinating between AI outputs and the core functions that actually run the business, like hiring, payroll, and closing the books.

The report, The copy/paste economy: Why task-oriented AI is failing the enterprise, is based on a survey of 200 professionals from Hong Kong and Taiwan across finance, HR, IT, and operations, conducted as part of a broader global study of 6,100 professionals – all active AI users at organizations with 500 or more employees.

These employees say they believe in their work, and they believe in AI – but they're stuck doing too much of the wrong kind of work. When AI is treated as a bolt-on tool for isolated tasks, it simply speeds up busywork.

The real transformation happens when AI is integrated into the systems and workflows that already run the business, so people can focus on the strategic, creative, and human work that truly moves organizations forward.

Employees are engaged, but buried in manual integration

Contrary to the prevailing narrative that employees are disengaged and fearful of AI, the research reveals that 97% of respondents rate their day-to-day work positively. Around 9 in 10 report a strong sense of progress, ownership, and connection to organizational goals.

But they are spending an enormous part of their day doing busywork that systems could be doing:

• 86% say their work requires them to reconcile conflicting data from different tools
• 75% say they spend significant time on admin and bureaucracy that creates friction in their day-to-day work
• Around a quarter lose more than seven hours a week to moving information and reconciling data

Before AI, people connected systems manually, moving data, chasing approvals, translating information from one tool to the next.

"My day feels busy but not genuinely productive when I'm pulled into constant coordination tasks and system-related issues that interrupt my actual work," says one study participant, a director-level employee in construction.

The research indicates that employees are not disengaged. They are overloaded with work that systems should be doing.

AI is more effective within workflows than as a peripheral tool

Most organizations have introduced AI, but only 21% have embedded it directly into core workflows – lower than the global average of 27%. The rest are using it only at the periphery for individual tasks like drafting emails, summarizing documents, and answering isolated questions.

AI agents operating without access to an organization's core systems tend to generate results that look reasonable, but they lack context and risk violating compliance rules because they cannot access the policies, approval chains, and data models that encode how a business actually runs.

Four in five employees in Hong Kong and Taiwan say missing or unclear information has delayed decisions, and 79% say their teams often disagree over whose numbers are right.

Compounding this, 74% of local employees report spending time redoing work because of business system or data issues – 10 percentage points higher than the global benchmark of 64%.

The difference in outcomes is stark. Among organizations in Asia Pacific and Japan with AI embedded in core systems, 65% of employees report time savings of 25% or more.

Where AI sits outside core systems, that figure drops to 36% – making employees in AI-embedded organizations 1.8 times more likely to report meaningful time savings.

Employees want to use AI for more high-impact work. The top use cases already in practice are:

• 52% monitoring metrics and suggesting actions
• 47% assisting with onboarding
• 46% routing cross-departmental approvals
• 41% budgeting and forecasting
• 41% answering HR and policy questions
• 40% supporting financial close or reporting

Even among organizations scaling AI in Asia Pacific and Japan, 4 in 5 employees still say navigating processes and systems is a source of stress.

AI adoption increases when the system is trustworthy 

Nearly 9 in 10 (89%) employees across Hong Kong and Taiwan say their confidence in AI increases when they trust the underlying system and data.

Employees already trust the systems they use to run payroll, close the books, and manage their teams. When AI operates inside those same systems, the confidence gap largely disappears.

However, that confidence gap is wider in Hong Kong and Taiwan than elsewhere. Only 69% of local employees feel their organization has a widely trusted source of truth, compared to a global benchmark of 78%.

Confidence in visibility over people and skills data stands at just 73% locally, against a global benchmark of 80%. These gaps in data trust represent both the root cause of friction and the clearest opportunity for improvement.

The report outlines a roadmap for leaders who want to move beyond task-based use cases and build an AI-powered operating model, including how to:

• Fix the foundation first. Modernize core systems and data so AI can produce consistent, reliable outputs.
• Integrate AI into end-to-end processes. Move from isolated task boosts to workflows where AI and humans each own clearly defined steps.
• Design for embedded, intelligent, invisible AI. Ensure AI is delivered in the flow of work, inside the tools and systems employees already use, rather than as yet another standalone solution.

Read the full research report.

Methodology: The research was conducted online globally by The Harris Poll on behalf of Workday among 6,100 global professionals (including 100 in Hong Kong and 100 in Taiwan) across HR, finance, IT and operations – all actively using AI at organizations with 500+ employees. The survey was conducted March 2-24, 2026.