Think of these as notes from the field, three-quarters of the way through 2026. As I’ve been helping companies work through AI transformation, these are five things I keep coming back to.

1. Uncertainty is the price of entry into AI transformation.

AI transformation is difficult, in part, because it's unpredictable. It’s uncertain. In the beginning, it's also unknowable. This is because AI transformation is, in many ways, more of a direction than a destination. It's both a process and a journey, and that journey looks different for every business. Transformation will change the reality of your business, but only if you take the first step, which is committing to the change, even though what you’re committing to isn’t exactly clear yet.

2. The leader has to change first.

If you’re the leader of the organization, you bear the ultimate responsibility for change. I don't just mean responsibility for the P&L or accountability to your board. I mean you have to take on the transformation for yourself, as an individual. You have to be the change you want to see in your organization. You have to get familiar with the tools, preach the gospel of AI transformation every chance you get, and be the most consistent supporter of the AI transformation you're asking everyone to undertake. Getting this right is one of the single biggest predictors of success in AI transformation.

3. Participation can’t be optional.

Once you and your company have committed to a direction, you can't make some people responsible while allowing other people to opt out. This is trickier than it seems because many people’s broader feelings about AI and what it means for their future can get wrapped up in the transformation you're trying to implement. The thing to remember is that people don't have to resolve every personal feeling they have about AI before participating in changes to the way the company works.

In the beginning, find the people willing and excited to help you move forward. Identify the champions and equip them, empower them, and support them. You won't always get this right. Someone you thought was perfect for the AI council may turn out to be a bad fit. That's okay. It just means you have to adjust. People can (and should) question decisions and point out what isn't working, but helping the organization move forward is a minimum expectation of everyone's job.

4. You need to embrace inefficiency for a period of time.

So much is unknown at the beginning of the process that waste and inefficiency are inevitable. That may cause frustration, but it's a necessary part of the work. Creativity is inefficient. Giving people time to experiment is inefficient. Finding the right mix of people for your AI council is inefficient. Exploring use cases that don't justify a pilot is inefficient. Much of this inefficiency is unavoidable, but that doesn't mean it isn't useful.

Each of these inefficiencies can teach you something useful about what you need to do next. Eventually, the work needs to produce evidence of value. In the beginning, though, judging every effort by how much time or money it immediately saves can prevent you from finding out what’s possible.

5. You have to be willing to work with imperfect solutions.

At some point, you have to decide whether a solution is useful enough to put to work, even though it isn’t everything you hoped it would be.

An AI-assisted workflow might still require a person to check the output or move information between systems. A first version might work for one specific task while leaving the rest of the process unchanged. Those limitations don’t necessarily make it a bad solution. It might be giving you something concrete to evaluate and improve.

You still have to understand what the solution can do reliably and where someone needs to intervene. Accepting imperfection doesn’t mean accepting careless work. However, it does mean recognizing that a useful first version can help you figure out what the next version should look like.

The work of transformation will ask you to make that judgment repeatedly. You have to be willing to put something useful to work and keep making it better. Perfect solutions sound enticing, but they rarely exist.

The Work Continues

These are my notes from the field as we head into the last few months of 2026. I expect I’ll have more to add, and I may come to see some of these things differently as the work continues. That’s part of transformation, too: letting what you learn change how you think. But for now, this is where my thinking stands.