Beyond Transformation

After years of process transformations and digital transformations, the word itself has become a warning. People hear “transformation” and they brace for disruption. They fear that this year's initiative will be unwound by next year's initiative. This fatigue is rational. Research consistently shows that most large-scale transformation programs fail to deliver their promised value.

Constant upheaval breeds cynicism. As John Kotter warns in Leading Change, without genuine buy-in, even the best strategies can be derailed when people pretend to cooperate while doing everything possible to undermine efforts they never believed in.

AI has intensified this dynamic. In 2016, AI pioneer Geoffrey Hinton declared that radiologists were “like the coyote that's already over the edge of the cliff” and advised stopping their training immediately. Nearly a decade later, AI has not come close to replacing radiologists. This prediction failed because it ignored organizational culture and how work actually gets done.

This creates a trust problem we should take seriously. One study shows that employee trust in company-provided AI has declined significantly, with trust in autonomous AI systems dropping 89%. In some companies nearly half of employees with access to approved AI tools use unauthorized alternatives instead. This signals resistance not to AI itself, but to how organizations are deploying it.

Transformation implies a bounded event with a beginning and an end. It suggests that once the process is complete, we will return to stability. But our operating environment will not be stabilizing. Our businesses shift, our technology evolves, and our client's expectations continuously increase. If we transform and then settle, we are simply accumulating the conditions for our next transformation.

The alternative is to build management systems capable of continuous adaptation. Processes that evolve steadily and constantly. This is not a transformation program with better branding. It is a fundamentally different operating philosophy.

In this model, AI becomes one component of our ongoing improvement capability, not the centerpiece of a dramatic overhaul. Our teams continuously identify opportunities, test solutions, and refine approaches. We become self-correcting, navigating disruption faster than our competition. Our goal is not a successful AI transformation. Our goal is to build an organization that no longer needs transformations.

Any change initiative that fails to answer “What's in it for me?” from the perspective of those expected to change is already beginning to fail. Sustainable change requires genuine alignment between our organizational objectives and individual interests.

The fear that AI will replace workers is legitimate. But the evidence suggests that the reality is more nuanced. Research shows that AI benefits those with the lowest initial ability the most. It turns poor performers into good ones, boosting the least creative, helping the weakest writers produce solid work. In one call center study, the lowest performing workers became 35% more productive with AI assistance, while experienced workers gained very little.

The implication: AI makes average workers excellent, but does not make them exceptional. Those with deep expertise developed through years of deliberate practice retain distinctive value. AI is a leveling force that raises the floor, not the ceiling.

Our message to our people needs to be clear: we are investing in you to thrive alongside AI, not replacing you with it. This means developing AI proficiency along side our distinctly human capabilities. We cannot replace judgment, empathy, creativity, or relationship building, with AI. When our people learn to collaborate effectively with AI, performance will improve beyond what either humans or machines achieve independently.

There is an old principle in process improvement, often termed “don't pave over the cow path”. I.E. don't automate a broken process. Doing so simply produces broken outputs faster. The same principle applies to AI. Layering artificial intelligence onto workflows that are unclear, inconsistent, or poorly understood does not create efficiency. It creates expensive confusion.

Before AI enters the picture, we need to understand our current processes deeply. How does work actually flow (not how it is documented)? Where is the value created and where does wasted time accumulate? What requires human judgment and what is truly routine. Lean principles can help guide us here. We can use them to map value streams and eliminate unnecessary steps. Providing standardization where appropriate.

Good AI requires good data. Data quality, clear definitions, and appropriate access are prerequisites for reliable outputs. We have to make data quality a prerequisite to implementation of AI automations.

The sequence matters: optimize the process, ensure data quality, then introduce AI where it adds genuine value. This requires thoughtful work on the front end, but it produces results that compound over time rather than fading away.

The future is human-AI collaboration where each contributes what it does best. In this model, AI serves as the worker, executing routine tasks, processing information at scale, generating drafts for review. The human serves as the supervisor, making final decisions, applying judgment to ambiguous situations, maintaining accountability for outcomes. But the human role extends further, to be the sentinel that identifies opportunities for continuous improvement.

This creates a self-sustaining feedback loop that transforms one-time implementation into ongoing capability. Our people closest to the work already know where their processes bog down. They will know where AI need refinement and where new applications could create value. When these insights flow into continuous improvement rather than accumulating until the next transformation we will become genuinely adaptive.

This requires designing AI systems with our people, not for them. Tools imposed without input feel mandated, not co-created. Workers are 72% more likely to trust AI when given interactive opportunities to practice and shape its use. We should treat our people as partners in the design, not subjects of the deployment.

Trust determines whether AI delivers value or becomes another failed initiative. People who trust that AI systems were built for them and will deliver real value use those systems more and use them better. Those who do not trust find workarounds or wait for the initiative to fade.

Building trust requires leading with conviction. We need to communicate a clear point of view on where we are going and why it matters. This requires creating genuine shared purpose that helps our people see their stake in the outcome and their role in achieving it.

Our team leaders are the critical layer. Employees trust direct managers more than the broader organization. When new tools arrive, workers look to their managers for guidance. If the leader explains why it matters and models its use, adoption feels natural. If the leader is skeptical or silent, adoption stalls regardless of executive enthusiasm.

We need both top down and bottom up change initiatives. We need to give leaders time to experiment, understand value, and develop conviction. We need an environment where experimentation is safe and progress matters more than perfection. One that our teams embrace and advance.

Success here looks different from traditional transformation metrics. We need to focus both on tangible organizational goals and people focused success.

The organization should be achieving significant efficiencies in how work gets done. We should have measurable cost savings for our clients, increases in our retention and win rates, a boost in profit and competitive advantages for our business.

But we also need a workforce that trusts the process, utilizes their most valuable skills while offloading low value work and sees continuous improvement as their responsibility rather than something imposed from above.

When our teams identify opportunities and implement improvements without waiting for directives and the system runs without perpetual top-down intervention, the required capability has taken root. This is our destination: not a transformed organization, but an adaptive one. Not a completed AI implementation, but an ongoing capability for AI-enabled improvement.