Every AI modernization plan eventually reaches the warehouse floor, the hospital ward, or the truck cab. That’s where it tends to stall.
Somewhere in your organization there’s a slide deck about AI. It has a roadmap, a budget, and a timeline. And somewhere else, far from the boardroom, there’s a six-year-old handheld scanner with a swelling battery that a driver reboots three times a shift just to clock in.
Those two things are connected, even though almost no one plans for them together. The strategy is bought at the top. It breaks at the bottom, at the edge, where the actual work happens and where the device either holds up or it doesn’t.
We’ve spent more than 19 years managing mobile devices for Canadian organizations in transportation, retail, healthcare, manufacturing, and the field. And the pattern we see is consistent. Companies pour money and ambition into transformation, then watch it sputter on the one layer they treated as an afterthought: ongoing management. If you run IT for an organization where the work happens on its feet, this is probably familiar.
The frustrating part is that the strategy is usually sound. The problem isn’t the vision, it’s the distance between the vision and the floor. And that distance is measured in devices, the physical things that have to be bought, configured, secured, supported, and eventually retired before any of the upstream investment touches a single frontline worker. That work is unglamorous. It’s also the difference between a transformation that ships and one that quietly doesn’t.
The workforce nobody designs technology for
Most enterprise software is built for someone sitting at a desk. The frontline worker, the nurse, the picker, the driver, the store associate, has largely been an afterthought, and the numbers behind that neglect are hard to ignore.
The venture firm Emergence Capital put it plainly: deskless workers are “more than 2.7 billion strong and representing 80% of workers worldwide,” yet for years a tiny fraction of software investment went their way. Emergence pegged it at less than 1% of business-software venture funding. The result is a workforce that runs the physical economy on tools designed for someone else’s job.
In Canada, this isn’t a niche. The frontline is the economy. Retail trade employs roughly 2.3 million people, the single largest employer in the country, and health care another 2.1 million. Add manufacturing, logistics, and the trades and you’re describing millions of Canadians whose productivity depends, every single shift, on whether a piece of mobile hardware works.
Here’s what actually happens when that hardware is an afterthought. The device becomes the bottleneck. The fancy new application can only be as good as the screen it runs on, the battery that powers it, and the network it holds. Underinvest in the edge and you’ve quietly capped the return on everything upstream of it.
Where AI modernization quietly stalls
Walk into any operation that runs on mobile devices and ask the IT lead what eats their week. It’s rarely the strategic work. It’s the broken scanner in aisle nine, the tablet that won’t enroll, the driver stranded because his device won’t connect at the dock.
Device problems are expensive in a way that hides from the budget. Research from VDC Research, summarized here by Getac, found that 75 to 95 percent of an enterprise device’s true cost lands after the purchase, in support, downtime, and lost productivity, and that consumer-grade devices fail two to three times more often than rugged ones built for the work. The sticker price is the part you see. The rest arrives quietly, one trouble ticket at a time.
And the support burden isn’t small. An Oxford Economics study found most companies have given up trying to carry it alone: 53 percent have fully outsourced mobile management, and another 34 percent outsource part of it. Only 13 percent still handle every piece in-house. That’s not a coincidence. It’s what happens when a small IT team realizes that chasing devices is no longer a job they can win by hiring one more person.
This is the layer where a modernization plan goes to die. Not in a dramatic failure, but in a slow drag: the rollout that slips because the old fleet can’t run the new software, the project that stalls because half the devices in the field are out of warranty and nobody owns the refresh.
And the drag compounds, because the device problem is almost never just one device. Here’s what actually happens in a fleet that’s been left to age: the failures don’t arrive evenly, they cluster. A batch of units bought in the same quarter hits end of life in the same quarter. Batteries degrade on roughly the same timeline. So the IT team doesn’t get a steady trickle of repairs they can plan around. They get a wave, usually at the worst possible moment, in peak season or mid-rollout. By then the model is often discontinued, replacement parts are scarce, and the “quick fix” of buying a few more of the same units is no longer on the table.
Rugged devices are built to last, five years, sometimes ten or more, which is exactly why the aging problem sneaks up on people. A fleet that still technically works can quietly fall years behind what the software it’s meant to run now demands. The device didn’t fail. It just stopped being enough, and nobody noticed until the new initiative arrived and asked it to do something it couldn’t.
AI modernization raises the bar on devices that were already behind
Now add AI to the picture, and the gap gets wider, not narrower.
The whole point of the AI conversation happening in Canadian boardrooms right now is to push intelligence closer to where decisions get made: predictive alerts on the warehouse floor, AI-assisted tools in the field, real-time capture and analysis in the moment. That ambition lands squarely on the edge device. It asks more of the processor, the camera, the battery, and the connection than the last generation of work ever did.
Industry analysts expect this to accelerate fast. Gartner forecasts that by 2029, at least 60 percent of edge computing deployments will use composite AI, up from less than 5 percent in 2023. Whether the exact number holds or not, the direction is unmistakable: the device in your worker’s hand is being asked to do more.
The trouble is timing. Most Canadian businesses in the frontline sectors haven’t started. Statistics Canada found that while national AI adoption doubled in a year to 12.2 percent, it sat at just 1.8 percent in transportation and warehousing, among the lowest of any industry. So the demand on edge devices is climbing at exactly the moment the fleets meant to carry it are oldest and the teams managing them are most stretched. Rising expectations, aging hardware. That’s the squeeze.
There’s a version of the next 24 months that’s easy to predict. The AI strategy gets approved. The pilot goes well on a handful of new devices in a controlled setting. Then it meets the real fleet, the one that’s been in trucks and warehouses for years, and the rollout slows to the speed of the slowest device. The lesson organizations tend to learn the expensive way is that you can’t bolt a 2026 ambition onto a 2020 fleet and expect it to hold.
What getting the edge right looks like
None of this means the answer is to buy newer devices and hope. We’ve watched organizations do exactly that, replace the hardware, change nothing about how it’s managed, and end up in the same place 18 months later. The fix isn’t a purchase. It’s treating the device as something with a full life that someone has to own, from the day it’s sourced to the day its data is destroyed.
A couple of examples from our own work make the point better than theory.
A major Canadian research hospital was trying to move its nurses onto a new electronic health record system, and the old handhelds simply couldn’t keep up: too slow, too crash-prone, increasingly incompatible with where the hospital was going. The modernization wasn’t blocked by the software or the strategy. It was blocked by the device. We sourced modern, scan-ready hardware built for clinical use so the move forward wasn’t held hostage by the thing in the nurse’s hand.
A national freight carrier had a fleet of roughly 800 mobile computers that were six years old and failing in the field, swelling batteries, constant downtime, no path forward. The easy sale would have been more of the same model. Instead we re-equipped the entire fleet with rugged devices, staged and deployed across depots from coast to coast in six weeks, and securely retired every legacy unit. Driver downtime dropped, and the break-fix burden came off an overstretched IT team entirely.
In both cases the technology that mattered wasn’t exotic. What mattered was that someone owned the whole lifecycle, sourcing, staging, security, support, and decommissioning, as one continuous job rather than five disconnected scrambles.
Cost is part of that picture too, and it’s where a complex fleet hides waste. When you’re running thousands of devices and lines across departments, locations, and carriers, almost no IT team has time to reconcile all of it. That’s the problem we built ClearSight TEMs AI to solve: it reads usage and cost across the whole fleet and turns it into a clear picture of what you have and where you can optimize, so you spend smarter, not just less. The waste was never about the carrier overcharging. It’s about complexity no human has the hours to untangle.
The part that doesn’t show up on a spec sheet
There’s one more reason the edge layer deserves more attention than it gets, and it has nothing to do with hardware specs. It’s about who is accountable when something goes wrong, and where your data lives when a device reaches the end of its life.
The stakes there are real and rising. The average data breach in Canada now costs CA$6.98 million, up more than 10 percent in a year, even as breach costs fell globally. A mobile device that’s lost, stolen, or retired without its data properly destroyed is a quiet entry point to exactly that kind of event. And the regulatory exposure isn’t uniform: federally, PIPEDA caps penalties at a relatively modest CA$100,000, but a single Quebec customer pulls your fleet into Law 25, where penalties can reach CA$25 million or 4 percent of worldwide turnover.
This is where being Canadian-operated stops being a slogan and starts being a control. Our service desk, our staging facilities, and our technical teams are all in Canada, and they’re our employees, not subcontractors. When a scanner goes down, a Canadian technician picks up. When a device is retired, it’s decommissioned with documented chain of custody. The people and facilities are here, and we control them. That’s what operational sovereignty actually means, and it’s the kind of thing you only appreciate the moment a device goes missing.
So here’s the reframe worth carrying out of all this. Managed mobility isn’t device logistics, and it isn’t a line item to minimize. It’s the layer that decides whether everything you’re investing in upstream, the AI, the connectivity, the modernization, ever actually reaches the people doing the work. Get the edge right and the rest of the plan has somewhere to land. Get it wrong and the best strategy in the building stalls in a warehouse aisle.
If your fleet is aging into a transformation it can’t yet carry, that’s worth a conversation before the next rollout, not after it stalls. And if the cost side is where the complexity feels worst, start small. Put a single month of fleet data through ClearSight and see what it surfaces.