I often hear the same types of complaints from leaders struggling with underperforming AI projects. Things started out great but then tailed off. They’re not seeing the ROI they expected. Maybe AI wasn’t a good fit for their organization.
These leaders are not alone. But whatever their specific underlying issue, I usually respond with the same question: How are you handling change management?
And the answer, in most cases, is not very well. It’s not that they’re not including any change management in their AI efforts. It’s just not getting the attention it should.
They’re using outdated methods and techniques. They’re treating change management as a commodity instead of a strategic advantage. They’re checking a box.
But that approach won’t work for AI.
AI is transformative by definition. It’s technology that makes change part of what organizations do. And when change is constant, change management needs to be constant as well. You can’t just run through the same standardized list of stakeholder interviews, communication plans and training regimens that you used for the last software rollout and hope for the best.
But what you can do, ironically enough, is use AI itself to reinvent change management for the age of AI by making your process more continuous, employee-centered and measurable. And in doing so, you’ll be transforming change management from a one-time project into an ongoing organizational capability.
Why traditional change management is breaking down
Here’s the paradox. When asked, every one of those leaders who told me about their AI issues would describe change management as essential. Yet when budgets get tight, it’s likely the first thing to be pared back. So how “essential” is it, really? If we thought it had real value, we wouldn’t be so quick to give up on it.
The truth is, change management hasn’t changed much over the past two decades. As a result, many of us are probably still relying on longstanding programs or disciplines that haven’t kept up with the pace of change. For example, we may be listening to the wrong people. Most programs still collect feedback through a handful of our colleagues and time-tested methods, including:
- Small interview groups
- Executive stakeholders
- Representative committees
- Broad but shallow surveys
The problem with this approach, particularly when it comes to AI, is that the people actually doing the work often have little voice. In other words, frontline employees are expected to change their behavior without helping to shape the solution.
Think about trying to use AI to redesign your sales function. If all of your change management effort is focused solely on the input of your sales operations team instead of your actual sellers, you’re likely to wind up with technology that looks good on paper but doesn’t fit real workflows.
And that’s why you see the familiar sparkle-and-fade patterns when it comes to AI adoption:
- Initial compliance, as everybody is willing to give the new tech a try.
- Post-launch celebration, as the organization sees early adoption as a sign of long-term success ahead.
- Steadily declining usage, as users begin to realize that AI doesn’t fit well with their work.
- A frantic after-the-fact scramble, as the organization attempts to diagnose problems and salvage its investment.
You can just hear the air coming out of the balloon as you read down that list. Listening to the wrong people – or not listening to enough of the right ones – is a recipe for discovering adoption barriers after the rollout instead of while the change is happening.
In most cases, that’s way too late.
How AI can reinvent change management
I can picture those leaders with their stalled AI projects shaking their heads at me. Even if we wanted to upgrade our change management process and listen to more employees, they might say, how are we supposed to do that?
It’s not a bad question. After all, they probably weren’t just being stubborn by not interviewing their thousands of employees for previous rollouts. It was simply a prohibitively expensive proposition.
But AI makes continuous listening possible. AI voice agents, for example, can conduct simultaneous conversations at enterprise scale.
The result? We can now work with organizations to compile feedback from hundreds of employees in under a week, a task that previously would have required months of interviews and surveys. AI gathered the information, and human change leaders interpreted it and made decisions.
AI, in other words, makes it possible to move beyond the limitations of traditional surveys and instead build an organizational nervous system. It allows for things like:
- Continuous employee listening through voice agents that enable open-ended conversations to yield rich sentiment and clear identification of pain points.
- Organizational network intelligence that uses data analytics to identify informal influencers, communication patterns, key adoption champions and disconnected teams that might need additional support.
- Behavioral monitoring and real-time feedback to identify where users get stuck, detect adoption friction early and recognize patterns before they become widespread problems.
For example, if sales reps consistently abandon an AI assistant before generating customer proposals, leaders can detect the pattern within days instead of discovering it months later through adoption reports.
These are the kinds of upgrades that can help organizations move from reactive change management to proactive adjustment. And with AI automating everything from data collection to synthesis to ongoing monitoring, experienced change leaders can use their expertise to focus on broader, more impactful activities like executive coaching and strategic communication.
In other words, less time and effort managing large administrative teams and more time influencing outcomes.
From change management project to continuous capability
I’m sure all of those leaders looking to reverse their slumping AI projects are probably thinking that while this approach to change management sounds great, what can they do about it right now? Well, Rome wasn’t built in a day, but those builders didn’t have AI. Here are three practical principles leaders can start using immediately:
1. Give every employee a voice
The old playbook of relying solely on representative groups doesn’t cut it anymore, particularly if you’re not getting feedback from those who will be directly affected by the change. At the end of the day, those are the groups that will actually be responsible for operating in a new system, so if you want it to stick, you need to hear their pain points.
Fortunately, getting input from everybody is possible now, so it’s time to create scalable mechanisms that continuously capture those frontline experiences.
2. Anchor every decision to measurable business outcomes
Start with your desired KPIs – whether sales effectiveness, marketing performance, customer satisfaction, productivity or employee experience – and then work backward. If a proposed change doesn’t support your chosen outcomes, reconsider it.
3. Embed change management into operations
For far too long, many organizations have made the mistake of treating change management as a parallel workstream or something tacked onto a project plan. Now it’s time to start integrating everything – listening, enablement, measurement and communications – to help make change management part of how the organization operates.
Of course, in order to use AI to make real progress in any of these areas, you have to have buy-in. And employees may initially distrust something like AI-driven listening. That’s why it’s important to focus on…
- Transparency about data use.
- Anonymity where appropriate.
- Personalized, contextual communications.
Most importantly, employees are more likely to come around on AI if they see that the organization is actually responding to what they’re sharing. Don’t tell them how AI makes things better, show them.
AI gives organizations the change management they’ve always wanted
For years, too many organizations neglected change management. Why? Because they could get away with it for many projects. But not AI.
AI demands a change management process that is continuous, measurable, adaptive, employee-centered and embedded in daily work. Ambitious, sure, but also impractical with yesterday’s technology.
But AI changes that equation.
It may seem ironic, but AI is what can enable change management to become what it was always intended to be. That is, an ongoing capability that continuously listens, learns, adapts and helps organizations evolve alongside their workforce.
And just in time for the massive AI changes that have arrived.