Transformation is a Process

~960 Words | ~4min Read

If we learned anything from agile transformations, it’s that transformations aren’t projects. They’re processes. And the integration of AI into how our organizations work will follow the same pattern.

Our mental model for ‘transformations’ matters because there’s real time pressure to adapt right now! Engineering departments need to learn and grow quickly. But it’s easy to miss that it takes the people time to do that! Three to six months for a team to build consistent, quality skills to leverage AI effectively is normal! And that’s just the beginning. The full organizational transformation takes years! The new ways of working actually landing and make a difference to the business is not a project. Much as we want it to be.

The mistake most organizations make is treating this like a project with a finish line. You don’t. You can’t. This complete renewal means more than simple re-training. You are talking about changing PEOPLE! Transformations are ongoing, because people don’t just flip a switch. But as the maturity of the transformation improves, the transformation’s needs change. This is a process, not a project.

So what does this mean for engineering leaders? It means you need to steward the “AI transformation” differently than you’d manage a typical initiative.

Start with your early adopters. Not everyone is ready to learn a new way of working at the same time. Start with the teams most eager to try AI, the ones who already see the opportunity. Set them up for success. Give them the space to learn, make mistakes, and figure out what works. Because here’s the thing: adoption is not the same as utilization. Your early adopters aren’t just using AI tools. They’re adopting AI into their processes, creatively applying it to unlock new possibilities. And when they do, they can become the teachers for the next wave.

This is where the relay race begins. Each successful pilot team creates momentum for the next. But that only works if you’ve structured it right.

Create the conditions for learning. A team can’t learn new behaviors in a vacuum. They need both a new theory of work, that is a clear understanding of how work changes with AI. Then they need actual practice and experience with that new way of working. This is where most transformations fail. Leaders announce the change, run some training, and expect people to figure it out. But encoding new behaviors takes repetition. Research suggests 21 or more repetitions to make a new pattern automatic. You can’t rely on the typical flow of work to provide that, after a 1-off training. A team might face a particular challenge only quarterly. You need to deliberately create practice opportunities so people can build the muscle.

Keep in mind, you need psychological safety! Remove the barriers and time pressures that make people afraid to surface mistakes. Teams that feel safe to fail. If they are still under performance pressure, don’t expect them to experiment, or to surface what they’re learning. Teams that intentionally slow down, are teams that learn faster. They iterate. They capture insights. They build systemic understanding instead of just following procedures.

Create the victory stories. Short-term wins matter. They build morale. They prove the concept works. But more importantly, they create the narrative that makes the next group want to adopt. When the first pilot team ships something meaningful with AI,that story spreads. When they solve a real business problem, it gets attention. When they demonstrate that this new way of working actually works, It encourages the next team want to try. It makes the organization want to invest more. You need an ongoing trickle of successes, to keep the transformation process fed!

Address the organizational bottlenecks in parallel. The transformation’s success depends on more than just the pilot team’s learning. It depends on the organization being able to accept and use what they produce. This about your value stream. Is your deployment pipeline is slow? Can your architecture can’t handle the new patterns the pilot team is discovering? Do your organizational structure create friction? Those bottlenecks will kill the momentum of your transformation. They sap the vital energy the trickle of successes creates. So while the team is learning, you have another job. You’re must identify and remove the obstacles that would otherwise block their progress. Obviously, you can’t do all this alone.

Build the guiding coalition. Transformations don’t succeed with one person pushing. You need leaders at multiple levels aligned on what you’re trying to do and why. This coalition doesn’t just support the pilot team. They also begin preparing the organization for the next wave. They’re thinking about succession. They think about how to scale what works. And they about what the organization needs to change to make this sustainable.

But here’s what you can’t do: declare victory too soon! The temptation is real. You get your first win, you celebrate, and then you’ll be tempted to think you’re done. You’re not. The new approaches are still fragile! The changes haven’t seeped into the organization’s culture yet. They can be undone. So instead, use the credibility you’ve gained to tackle bigger problems.

Use it to go after the systems and structures that still need to change. Use it to think about who you’re promoting, who you’re hiring, how you’re developing people. Because the next generation of leaders will determine whether this transformation actually sticks.

This is the work of years, not months. But it’s also the only way transformations actually work. Not as projects with finish lines. Transformation is a process. Each phase builds on the last. Early adopters become teachers. And your organization gradually becomes capable of doing things it couldn’t do before. The question is whether you’re willing to do the patient, deliberate work of actually transforming.