The Cost of Switching: An AI Lock-In Experiment (Post 5 of 6)

The Organizational Dimension: Shadow IT and Inertia

Up to this point, I’ve focused on my personal switching journey, what it felt like to try to move from ChatGPT to Claude or Gemini. But the bigger lesson is what happens when you zoom out from one individual to an entire organization.

Here’s the reality: most companies didn’t lead on AI adoption. Workers did.

The data backs it up: 45% of companies don’t pay for any AI tools, and 42% haven’t set clear usage policies . So what happens? Employees experiment on their own. They pay out of pocket. They sign up for free accounts. They create grassroots adoption patterns. In other words… shadow IT.

The AI Adoption Paradox points out the long-term consequence: by the time leadership finally decides to roll out an “official” AI strategy, they’re not introducing something new to a blank slate. They’re trying to dislodge deeply ingrained personal workflows . Every employee has already built habits, shortcuts, and trust with their preferred tool. At that point, leadership isn’t just introducing a new technology, they’re asking people to switch.

And switching is hard. I felt the friction myself as one person. Multiply that across hundreds or thousands of employees, and you start to see why organizations face massive inertia.

That’s why AI adoption at scale isn’t just about licenses or training. It’s about change management, navigating the cognitive, procedural, and trust-based switching costs that employees have already absorbed. Without a strategy, organizations inherit the chaos of grassroots lock-in.

If your company announced tomorrow that everyone had to switch to a single “official” AI tool, would your team embrace it or would it feel like starting over from scratch?