Your Employees Aren't Sabotaging Your AI Strategy—They're Telling You It Doesn't Work

1 day ago 4

Dan Adika is CEO & Co-Founder of WalkMe, the world’s first Digital Adoption Platform to help navigate technology change.

The story enterprise leadership is telling itself about AI right now goes something like this: "We gave them the tools and ran the training, yet a third of them are actively sabotaging the rollout, refusing to use what we deployed, pasting customer data into unapproved apps and gaming the metrics so AI looks worse than it is. The employees are the problem. If we can just get past the resistance, the strategy will work."

I've read that story in four different publications this month. I don't buy it.

I spent 15 years building a company that exists specifically to close the gap between the software enterprises buy and the way people actually use it. I've sat in rooms with the CIOs of some of the largest organizations in the world, and I can tell you that when close to half of your workforce is quietly going around your AI strategy, you do not have a sabotage problem. You have a product problem. Your people are not your adversaries. They are the most honest user feedback you will ever get, and they are trying to tell you something.

My company's fifth annual State of Digital Adoption Report, drawn from 3,750 executives and workers, found that 45% of employees have used unapproved AI tools in the last 30 days, and 36% have used them with confidential company, customer or employee data. Furthermore, 88% of executives believe they provide their teams with adequate AI tools. Only 21% of employees agree. And while 61% of executives trust AI for complex work, only 9% of employees do.

Employees are losing seven to nine hours a week to friction with the software they were given. That's 51 workdays a year, every year. This is not resistance. This is a workforce making a rational call. The sanctioned tool does not work well enough to do the job, but the unsanctioned one does. Workers have a deadline, so they choose the option that works. I imagine that every single person reading this would make the same choice.

The reason this reads as sabotage to leadership is that leadership is looking at the wrong dashboard. CEOs report AI progress to their boards based on license deployment, sanctioned-tool usage and internal survey data. Those metrics describe an enterprise that does not exist. The real enterprise is the one where 45% of your workforce is keeping your company running on tools you didn't buy. That enterprise shows up nowhere in the report.

That is the fundamental mistake of the current moment. Picture the sales rep on a team of 200 for which your company bought Copilot licenses. She's not resisting AI. She's drafting her follow-ups in ChatGPT on her phone because Copilot can't see her pipeline the way she needs it to. The finance lead isn't resisting AI either. He's running his variance analysis in Claude because the sanctioned tool takes four prompts to do what the unsanctioned one does in one. They already adopted AI months ago. They did it without you because waiting for you was going to cost them their quarter.

So, what should leadership actually do?

First, stop framing this as a workforce problem. Every time a CEO takes the stage and says, "Our people need to embrace AI," they are describing a company where the executives are the last ones to find out that their people already did. The language needs to change. The diagnosis needs to change.

Second, measure use rather than deployment. Deployment is what you paid for. Use is what you got. The number of employees logged into your sanctioned AI platform weekly is not a metric. The number of employees who finished a real piece of work using it versus using something else is the metric. If you can't answer that question, you can't manage this.

Third, find the 45%. They are telling you which tools work and which tools don't, not in a focus group or survey but in the most credible feedback mechanism there is, which is what people actually do when they have a deadline and nobody is watching. That is not a governance failure to be stamped out. That is the most detailed product research your company has ever conducted. The problem is that almost none of it reaches the executives who need to see it.

Building the visibility to surface this takes work, but it's necessary. It's time for leaders to stop calling their own employees saboteurs long enough to ask what they're trying to tell them.

Forbes Technology Council is an invitation-only community for world-class CIOs, CTOs and technology executives. Do I qualify?

Read Entire Article