How device performance is undermining employee productivity

IT budgets are growing faster in 2026 than they have in years, and AI is absorbing most of that growth. IT leaders now expect 2026 budgets to grow 5.8% year-over-year, and more of them are increasing spending on AI and machine learning than on any other category, even as they keep expanding device fleets and their stack of security agents. Every one of these purchases gets justified by the same pitch of faster work, fewer bottlenecks, and a rise in output.

In many organizations, that payoff never materializes, and the tools aren't to blame. The devices running underneath all of them are failing constantly, in ways IT dashboards don't track.

No ticket, no log, no idea it happened

A crash generates a signal. The app closes, the OS flags an error, or maybe the machine reboots. Somewhere, a log entry gets written. If it happens enough, a support ticket is opened. IT can see it, count it, and eventually justify budget to fix it.

A hang doesn't do any of that. We’ve all experienced moments sitting in front of a computer, staring at a frozen screen or an endlessly spinning cursor. The device is technically still running, but nothing crashed, nothing rebooted, and nothing produced an alert. All you can do is wait, then force-quit or restart, and move on with the day. No ticket. No log. Only anecdotal.

The failure mode causing the most damage is structurally invisible to the systems built to catch failure modes. Task-switching research backs this up. A study published in APA's Journal of Experimental Psychology  found that switching back to a task after an interruption costs real time beyond the interruption itself, a cost that grows the more complex or unfamiliar the task is.

More is being asked of the same unstable devices

Two unlogged hang ups per employee per day, across a workforce of 5,000 people, amounts to thousands of hours of lost productivity every single week. The same devices producing these numbers are now being asked to do more. AI adoption grew nearly 1,000% year over year, according to the Omnissa State of Digital Workspace.

The same devices struggling with these failures are now being asked to do more. AI adoption inside the enterprise grew roughly tenfold year over year, according to the Omnissa State of Digital Workspace 2026 report, and it isn't just Copilot driving that number.

On enterprise-managed iOS devices, employees run 96% of their AI usage unsanctioned, mostly through ChatGPT instead of the officially provisioned Copilot. On Android, that share climbs to 99%, led by Gemini. Whatever assistant is fastest to open gets used.

Hemant Sahani, Omnissa Vice President of Product Management for Workspace ONE, explained why on the AI at Work podcast. Most employees already experienced how much more productive these tools made their personal lives long before ChatGPT ever showed up on a work device. When the sanctioned option is slower to open than the one they already trust, the familiar tool wins.

A newer layer is stacking on top of that, too. Background agents are AI processes that keep running on their own, monitoring an inbox or drafting a summary without anyone actively prompting them. That load lands on devices that were already carrying a lot.

The Omnissa report also found that IT administration, security, and system management software, the tools that keep a device patched and monitored, already make up a large share of what's installed on the average enterprise machine. AI now competes with all of it for the same memory and processing power. Unlike opening and closing a single app, this new load runs continuously, leaving less room for a struggling system to recover on its own.

That load doesn't land evenly. The report's telemetry shows Windows devices failing more often than Macs across every category it tracked: 3.1x more forced shutdowns, 2.2x more app crashes, and 7.5x more app hangs.

Hangs show the widest platform gap of all. Windows produces 7.5x more of them than Mac. That's the failure type this piece has already argued costs the most, and it's also where the gap between platforms is largest.

That gap compounds through scale as well as severity. Windows holds a far larger share of the desktop market overall, roughly 62% to Mac's 15% globally. Enterprise fleets skew even further toward Windows, anchored by the legacy software and management tools most large organizations still depend on.

That means the platform with the least visible failure type also runs under the largest number of employees. A hang rate of 7.5x higher than Mac's isn't a narrow platform quirk. This is where this invisible cost actually piles up, because more of the workforce sits on it. The cost isn't smaller on Windows. Instead, it is spread across more people, and harder to find.

Companies are investing in AI expecting a productivity multiplier. If the hardware running that workload freezes instead of keeping up, part of the expected return gets absorbed by the same instability as those numbers describe. The ROI math behind every AI rollout assumes a stable foundation most fleets don't currently have.

In healthcare, the tax is paid in seconds

These delays compound fastest in clinical settings, where the currency isn't employee satisfaction. It's time, measured in seconds, at the bedside.

A clinician moving between exam rooms doesn't experience a slow login as a minor annoyance. What they experience is a spinning cursor standing between them and the next patient, while a room down the hall is already full. A frozen badge tap, a delayed single sign-on prompt, a virtual desktop session that takes an extra 15 seconds to load. None of that generates a ticket. All of it compounds across a shift built around dozens of patient encounters.

One health system in the northeastern U.S., serving more than 27,000 patients a day, runs Epic, the electronic health record (EHR) system clinicians chart in. When it moved clinicians onto virtual desktops to access Epic, the biggest gains had nothing to do with the EHR software itself.

When it moved clinicians onto virtual desktops to access Epic, the biggest gains had nothing to do with the EHR software itself. They came from stripping out what sat in front of it: eliminating the "double hop" between the desktop and Epic and integrating the Epic Hyperdrive directly into the virtual environment so clinicians could move between locations and pick back up where they left off.

All of it determines whether a physician finishes a shift on time or behind, and whether the patient in the next room gets seen before a manageable issue turns into something more serious.

Physicians don't evaluate a device the way an office worker might. They don't particularly care what platform it runs, or which vendor built the login tool. They care whether it loads fast enough and stays working long enough to get them to the next patient. Health systems have spent years treating patient experience and clinician burnout as two separate problems.

What “digital experience” means for retail

Healthcare treats digital experience as inseparable from patient care, valuable in its own right. Retail sits on the opposite end of that spectrum. An associate's device experience matters less on its own terms and more for what it does to the customer standing in front of them.

Retail runs on razor-thin margins and a competitive landscape where customers have no shortage of options. If a shopper can't find what they need, or an item is out of stock, the alternative isn't a phone call to corporate. It's a five-minute drive to a competitor.

That's the lens retailers are applying to their device fleets, and it explains some of the specific choices the report surfaces. A large home improvement retailer, for example, equips store associates with ruggedized handhelds that show real-time inventory and help locate items on the shelf.

That's not about making the associate's day nicer. A customer stuck in an aisle for three extra minutes might just leave. AI-driven dynamic pricing and store-flow optimization get adopted for the same underlying reason, to shave seconds off a process where the competitor is one parking lot away.

AI-driven dynamic pricing and store-flow optimization are being adopted for the same underlying reason: to shave seconds and friction off a process where the competitor is one parking lot away.

How organizations should be measuring experience

This pattern isn't limited to healthcare and retail. In any business, digital employee experience (DEX) isn't just a soft HR metric. Follow it far enough, and it becomes something concrete: a clinician's ability to move through a shift, a shopper's decision to stay in the store or leave, or how much output a company gets back for what it already pays its people.

The report's DEX composite score proves the point. It's built from measurable hardware, software, and stability data, not a sentiment survey, and it reveals a real, quantifiable divide between platforms. A 120% higher "Good" DEX score for Mac over Windows isn't a preference. It's the direct result of fewer crashes, fewer forced shutdowns, and far fewer hangs.

So, who pulls ahead in 2026? Not the organizations buying the most additional tooling. It'll be the ones that finally gain visibility into the failure mode hiding in plain sight. It never generates a ticket. It** never shows up in a budget line. And it quietly determines whether your investments in AI, device fleets, and the people who use them actually pay off.

See what your tickets aren't telling you

Your employees won't always raise their hand when work slows down. Omnissa Workspace ONE Experience Management gives IT a clear view of device, app, and network performance across desktop, mobile, and virtual endpoints. Experience scores, AI-driven insights, and root cause analysis help you spot hidden friction and fix it before it costs you.

Ready to turn DEX into a measurable business advantage? Explore the Omnissa platform or connect with your Omnissa team to see DEX in action.