Employee Experience Doesn't Break. It Drifts.

Most leaders think employee experience is a state. Good or bad. Healthy or toxic. A number on an engagement survey that either clears the bar or doesn't.

That's the wrong model. Employee experience is not a state. It's a slope.

It drifts. Slowly, quietly, and mostly out of view — until one day the work is worse, the people are tired, and nobody can point to the moment it went sideways. There wasn't a moment. There was a slide. And the customer meets the bottom of that slide long after the employees started sliding down it.

Even Disney drifted

Here's the part that should worry anyone confident their brand is too strong to slip.

In the mid-1970s, Disneyland's customer satisfaction scores dropped. Disney — the company whose entire reputation is built on a controlled, magical, repeatable experience — watched the numbers fall and went looking for the cause. What they found wasn't a marketing problem or a pricing problem. It was an employee problem.

Morale had deteriorated. Staff no longer felt their jobs had any real connection to delivering the Disney experience. Departments had turned on each other in cross-functional feuds. The people responsible for the magic had lost the thread that tied their daily work to why it mattered.

The repair is the interesting part. Disney didn't fix it with a slogan or a new incentive scheme. They rebuilt training so it went beyond task instruction and restored an emotional connection between the employee and the experience they were there to create. They empowered people to go beyond their formal job descriptions — to do the extra, unscripted things that make an experience feel human — when the moment called for it. The park recovered. It went on to serve more than 16 million visitors a year.

The lesson isn't "Disney is magical." The lesson is that even a company organized entirely around experience will drift when employees lose their felt connection to the work — and that the drift shows up inside the building before it shows up at the gate.

What drift actually is

Let me define it precisely, because "morale is down" is too vague to act on.

Employee experience drifts when the story leadership tells about the work stops matching the conditions under which employees actually do it.

That gap is the whole thing. Leadership says one story. The workload tells another. And every day the two stories fail to reconcile, the drift widens.

I studied this directly. My dissertation looked at middle managers being asked to deliver customer-centric change while the organization around them quietly made that impossible. The pattern was almost mechanical. Leadership talked about the customer constantly. But nobody had translated that vision into anything the managers could act on. One participant put it plainly: the talk is customer, "but I don't know that I've ever heard anybody define what that means." Another couldn't recall whether her own department even had customer goals — "the fact that I can't remember should tell you that it's probably not in there."

Meanwhile the actual conditions were pulling the other way:

• Layoffs. Wave after wave of workforce reductions, with the same work spread across fewer people. "We're doing it with fewer people and I don't think folks do their best work, I know I don't when I'm just tired out of my mind."

• Bureaucracy. A CEO promising to eliminate it while employees saw no evidence: "Here's a vision, here's the theory, here's the whatever, without any… I'm not seeing a lot of tangible stuff that people can actually act on."

• Legacy metrics. People asked to behave customer-first while still measured on the old transactional numbers — caught, as one framing put it, in the gap between the aspiration of the future and the structure of the past.

• Insight that went nowhere. Employees who gathered customer feedback and watched it evaporate: "I got all this stuff, but you don't do anything with it."

The emotional fingerprint was unmistakable. When participants drew their relationship to the organization, one word appeared in every single drawing: frustrated. Right behind it — tired, tense, distressed, sad, annoyed.

Now here's the detail that matters most. When those same people drew their relationship to the customer, the words changed completely: excited, delighted, happy. They wanted to help. They lit up at the idea of making someone's experience better. The care was never the problem.

That's the signature of drift. Not people who stopped caring. People who still cared, trapped inside an operating model that made caring pointless. The frustration wasn't aimed at customers. It was aimed at the gap.

AI is now the fastest way to widen that gap

Which brings us to the newest, most efficient drift engine available to leadership: the rushed AI rollout.

Look at what's happening in manufacturing. In Grant Thornton's 2026 AI Impact Survey, not one of the 100 manufacturing leaders reported a significant revenue increase from AI. Not one reported significant cost savings either. Across other industries in the same study, 12% reported each. A clean zero in a sample that size isn't noise. It's a signal.

It's not for lack of trying. Manufacturers are adopting aggressively — 64% cite efficiency gains, operations is where they most want more AI. But 48% are stuck in pilots, against 34% across industries. And when you ask why they're spending, 45% point to competitive pressure. Not a costed bottleneck. Not a defect rate they're chasing. The worry that a rival is a step ahead.

The Forbes analysis of that data lands the diagnosis hard: these companies bought the anxiety, not the problem. They purchase tools and wait for the vendor to figure out deployment. Pilots never resolve because nobody owns a number — "a pilot with no P&L target can't really succeed or fail. It can only continue." The broader MIT research says the same thing from a different angle: roughly 95% of enterprise generative-AI pilots showed no measurable P&L impact.

Everyone's reading this as an investment story. Weak ROI. Procurement indiscipline. Show me the number, show me the owner, show me when we stop.

They're missing half of it.

Pilot purgatory is an employee experience problem

Every one of those stuck pilots has people inside it.

When leadership buys AI out of competitive anxiety instead of a defined problem, employees inherit the ambiguity. They get handed a tool, a vague mandate to "transform," and a rollout with no stated success condition — which means no possible finish line. They're asked to change how they work, absorb the learning curve, keep every existing deliverable, and somehow produce proof of value for a project that was never scoped to produce it.

That is the exact structure of drift, rebuilt with newer parts. The story says transformation. The conditions say churn. New expectations pile on. Clarity, support, and meaning don't. And the same emotional fingerprint from my dissertation starts forming at desks running an AI pilot that will apparently never end.

Notice the parallel that should make any leader uncomfortable. My managers were asked to be customer-centric while measured on legacy numbers. Today's employees are asked to be AI-forward while measured on everything they were measured on before — plus proof, plus adoption, plus speed. Same trap. Newer tooling. The old deliverables never left; the new expectations just landed on top of them.

This is the core of it, and I want to be exact about the claim: People can survive hard change. What erodes them is incoherent change. My participants weren't allergic to effort or even to transformation — they got excited about improving the customer experience. What ground them down was change that arrived disconnected from meaning, proof, and support. AI rollouts built on anxiety deliver precisely that, at scale, faster than anything before them.

The warning

This is not an argument against AI. Disney's fix wasn't nostalgia, and the answer here isn't to refuse the tools. The manufacturers who close the gap won't be the ones holding out for a smarter model. They'll be the ones who change how they buy — start from a costed problem, name the metric, put one executive's name against it, and set a date it dies if it doesn't work.

Do that, and AI can become the kind of change people rally around, the way Disney's employees eventually rallied around a role they could finally connect to. Skip it, and you don't just waste capital. You teach your best people that transformation means enduring incoherence indefinitely.

So here's the part leadership keeps learning too late. By the time the returns fail to show up on the P&L, the drift has usually been showing up in the employee experience for months. The zero in that survey didn't appear the day the report was published. It was forming at every desk that got a tool and no reason, a mandate and no metric, a pilot and no end.

The dashboard is the last place drift shows up. Your people are the first. The only real question is whether you'll ask them before the number forces you to.

This is the third post in the series. The first laid out the three AI UX patterns and why "AI slop" is a signal. The second argued that your employees meet the slop first. This one is about what happens to them while they wait for the rest of the org to catch up.