Are you feeling the pressure to “do something” with AI?

Are you “doing something” with AI? Most leaders I’m talking to can point to visible AI activity inside their organizations, whether that’s a pilot underway or teams wiring small automations into their daily work. In many cases, people genuinely are feeling more efficient!

So why does it still feel unclear whether anything meaningful has changed for the business?

One reason is that AI has become unusually easy to touch. Anyone can open a tool, write a prompt, and get a result that looks polished. But there’s a risk that this feeling of motion becomes a proxy for progress. We celebrate time saved in a single step, even if the value chain as a whole hasn’t improved.

I see this most clearly in software delivery.

There’s definitely value in using AI to write code, generate tests, or help with refactoring. But if you zoom out, the biggest constraint in most organizations is clarity:

• About the problem worth solving
• About the tradeoffs leaders are willing to make
• About how a change will ripple through operations, customers, and adjacent systems.

If those questions remain unresolved, making code generation faster really just compresses the middle of the process.

When you step back and ask what actually changed for customers or margin or risk, the answer can be hard to identify.

To me, that’s a sign that the initiative began in the wrong place.

When an effort starts with “we need to do something with AI,” the conversation naturally turns toward where a tool can be applied, rather than where the business is actually stuck.

Here’s what I suggest:

1. Name a concrete outcome you care about, whether that’s shortening time to value for a new customer, improving reliability in a critical operation, or increasing the capacity of a team without burning people out.
2. Then look carefully at the value chain that produces that outcome today. Where does work pile up? Where does rework appear? Where are decisions delayed?
3. Only then should you ask whether AI is the right lever.

Sometimes it is. Other times, the better move is a change in process or policy instead.

If you’re feeling both the pressure to be doing something with AI and the responsibility to deliver clear business results, it may be worth pausing to ask:

• What result are we trying to change?
• Where, in the actual flow of work, is that result getting slowed down?
• And is AI the tool we need, or just the easiest to reach for?

If that’s something you’re working through right now, I'm always open to a conversation.

Talk soon,

Josh