For one in five enterprise workers, a full workday each week is not spent on actual work. It is spent moving data between systems that can't agree.As enterprises race to get value from AI, new research suggests that the systems supporting it haven't kept pace. A global study of 6,100 enterprise decision-makers, conducted by The Harris Poll on behalf of Workday, found widespread friction across core business workflows. When asked what enterprise leaders spend significant time doing, 81% say moving information between systems, and around three-quarters say the same about reconciling conflicting data or reports. One in five loses at least seven hours a week to moving information alone — and a similar share says the same about reconciling data.Yet 97% still rate their day-to-day work experience positively — the issue isn’t morale, it’s infrastructure.The pattern holds globally. Some 43% of employees report busy-but-unproductive days often or very often, but the burden is unevenly distributed. In the UK, that figure rises to nearly 60%, seventeen points above the global average. In Australia, 90% spend a significant amount of time moving information between systems — the highest rate among those surveyed. In the U.S., where enterprise system investment runs deepest, 85% report the same.Read on to see how this challenge takes shape across functions and industries — and how leading companies are responding.Where The Work Breaks DownAcross every business function, the same problem takes a different shape. In HR, 70% of respondents report spending significant time redoing work due to system issues — the highest rework burden of any function. Almost three in 10 HR professionals say that AI layered onto existing systems makes the problem worse, not better.In finance, the issue isn’t speed — it’s impact. One in five finance professionals says AI has accelerated their work without actually improving results, the highest rate of any function.In IT, the mood is mixed: governance and approval checks are the top barrier for 30% of professionals, yet 36% say AI has made their day meaningfully better.In operations, just 20% have AI built into core systems, the lowest of any function, yet 70% say AI has reduced their time on tasks, the highest of any function. Even with limited integration, the gains are already visible.The same problem manifests differently across sectors. In North America, hospitality respondents report especially high levels of data reconciliation and re-entry work, while healthcare respondents are more likely than the regional average to say AI has reduced task time and accelerated work in a productive way. Banking shows a different kind of constraint: governance and approval checks remain a visible barrier to realizing AI value. Across sectors, the pattern is less about whether AI is present and more about whether it is connected to the trusted systems where work actually happens.When organizations begin to address their system constraints, the results follow quickly. In HR, for example, rather than manually running reports to find data gaps, Workday Payroll Agent surfaces them instantly. “It reduces manual processes, drives compliance and easily assists with data insights,” says Monica Seiter, the director of payroll at Lindenwood University.AI Isn’t The Problem. The Systems Running It Are.Ask a business leader if they trust their AI, and most will say yes. Ask if they trust the data feeding it, and the answer gets complicated. A forecast built on incomplete inputs. A hiring decision shaped by records that haven't been updated. A contract reviewed without visibility into what changed last quarter. This is the data conundrum enterprises are living with — and 87% of decision-makers say AI only increases their confidence when they trust the underlying systems and data. AI models are becoming increasingly capable, but without the clean, connected data to support them, value cannot materialize. The barriers holding organizations back are often structural: governance and approval checks are the top barrier for 30% of IT professionals. Rigid systems are cited by 26% of HR respondents and 25% of finance respondents. In operations, data quality ranks highest at 27%.Where companies have addressed those constraints, the results are measurable. When U.S. tariff policy shifted rapidly, energy storage company FlexGen used Workday Contract Intelligence to automatically analyze exposure across its supplier agreements, identifying more than $50 million in tariff risk and avoiding over $35 million in potential costs. "[It] helped us understand exactly which agreements allowed us to pass costs through, which ones required notice and where our exposure was," says Anthony Tacker, FlexGen’s director of risk and procurement.At the healthcare supply chain company Vizient, Workday Revenue Contract Agent has transformed the way the company manages contracts. "It doesn't just extract and populate Workday — it understands the contract," says Megan Shaw Previte, senior director of accounting professional services revenue. "It's like having an intelligent accounting partner built into our workflow."Most Companies Have AI. Few Are Getting Much From It. Most enterprises have deployed AI. Far fewer are getting value from it. The gap comes down to one thing: how deeply AI is embedded in the systems where work actually happens. Right now, only 27% say it is — and a quarter say that layering AI onto existing systems increases friction rather than reducing it. Among organizations with AI deeply embedded in core systems, 60% report task time reductions of 25% or more, compared with 36% among organizations where AI is not used in core systems — about 1.7x higher. The same technology, built on different foundations, delivers nearly twice the productivity gain.Japan offers the clearest real-world picture of what happens when AI is fully embedded in how a business operates. With 39% of organizations running AI inside their core systems, the highest of any country surveyed and well above the 27% global average, Japan shows what is possible with best practices. Nearly three in four say AI has reduced their task completion time, compared with just over half globally. When AI is connected to core systems, 72.5% say it reduces the manual work burden, compared with 43% globally. The same holds at the company level. Global real estate firm JLL, managing 1.5 million job applications a year, embedded AI into its hiring process through HiredScore AI for Recruiting. The result: a 70% decrease in time to screen candidates, a fivefold increase in recruiter capacity and $12 million in projected headcount savings.For Snowflake's chief accounting officer, Emily Ho, having built a unified foundation across finance and HR, the next horizon is already visible. "Being able to have this system of record and deploy agents that can go across different modules on a single platform will be super powerful," she says.From AI Adoption To Impact: The Companies Closing The Gap By now, most business leaders have heard enough about AI to last a lifetime. The promises are familiar. The pilots are everywhere. The question that actually matters is harder to answer: not whether AI works, but when it will start changing how the business operates. For a growing number of organizations, that moment is already here. Not because they found a better AI solution, but because they built the right foundation underneath it. The shift looks different depending on where you sit. For some, it shows up in hiring. For others, it's contracts, compliance or how fast finance can close the books. But the pattern is consistent: when AI is connected to clean, unified data across the whole business, the manual work that consumes entire workdays starts to disappear.Chipotle cut time-to-hire by 75% and now moves candidates from application to start date in four days rather than 12.7-Eleven saves more than two million hours annually, with 85% of job applicants scheduled for interviews in under an hour.NetApp has analyzed 90,000 contracts using AI, saving more than $2.5 million and thousands of hours across legal and business operations.These aren't experiments. This is what AI looks like when it's built into how a business actually runs.When the operational load lifts, the role of the human isn’t diminished; it becomes more important. As Snowflake's Emily Ho puts it, "While AI can get tasks done, it cannot inspire. It cannot lead. It cannot make decisions that change people's lives."That's the goal: AI eliminating repetitive, manual work, allowing people to get back to the work that needs human judgment, creativity and leadership.