I don’t think the biggest opportunity with AI is the technology itself.
I think it’s understanding work well enough to know where AI can actually help.
Years ago, I was part of an initiative where my team thought we knew exactly where the work needed to begin.
Then we slowed down.
Instead of starting with the solution, we spent time understanding how the work actually happened. We talked with the people closest to it. Not because they had the technology answers, but because they understood where the work broke down, where it relied on unwritten knowledge, and where teams had quietly adapted just to keep things moving.
We ended up solving a different problem than the one we thought we had.
Looking back, one lesson has stayed with me.
Understanding the work didn’t just change the answer. It changed the question we should have been asking.
That lesson has stayed with me for years, but it has taken on new meaning as conversations about AI have accelerated.
Today, leaders are navigating an almost overwhelming number of possibilities. Every week introduces another model, another platform, another promise of transformation. The pace is exciting, but it also creates pressure to act quickly. It is understandable that many organizations begin by asking, What can AI do for us?
I wonder if there is a more useful first question.
Understanding work before choosing technology
When our team slowed down, we weren’t delaying progress. We were improving our understanding of where progress was actually needed.
At first, our assumptions seemed perfectly reasonable. We had experienced leaders, a clear objective, and a logical place to begin. Nothing about our thinking felt careless.
What changed wasn’t the initiative.
What changed was our understanding of the work.
As we spent time with the people closest to it, we realized the beginning of the process wasn’t where we thought it was. Important context existed long before the work became visible to leadership. Decisions were being shaped by conversations, judgment, and experience that never appeared on a process map.
The technology hadn’t changed.
The work hadn’t changed.
Our understanding had.
That distinction continues to shape how I think about AI today.
Why understanding the work matters
Most organizations already have process documentation.
They have
- Workflows.
- Operating procedures.
- System diagrams.
- Governance.
Those things matter.
But documented processes and lived work are not always the same thing.
The people doing the work understand where information arrives too late, where approvals create unnecessary delays, where one team compensates for another, and where informal workarounds have quietly become part of everyday operations.
Those workarounds are often treated as problems to eliminate.
Sometimes they are.
Sometimes they’re also signals.
They reveal where the formal process no longer reflects reality.
They reveal where employees have already adapted because the work demanded it.
Before asking whether AI can automate something, it is worth asking why that workaround exists in the first place.
Understanding that answer often changes the conversation.
The order of the questions matters
When organizations begin with technology, they naturally explore capabilities.
They compare platforms.
They evaluate vendors.
They benchmark what peers are trying.
Those conversations matter.
They should.
Outside perspectives expand what leaders believe is possible. Vendors introduce capabilities that internal teams may never have considered. Advisors recognize patterns across organizations.
The question isn’t whether outside expertise is valuable.
It absolutely is.
The question is what problem we bring to those conversations.
A vague problem invites an impressive demonstration.
A well-defined problem invites a meaningful discussion.
That is a very different starting point.
The people closest to the work aren’t choosing the strategy
One concern I sometimes hear is that listening to employees means allowing everyone to decide what should happen next.
That isn’t what I’m suggesting.
The people closest to the work are not responsible for setting enterprise priorities.
Leaders are.
What those employees often contribute is something equally valuable.
They help define reality.
They understand where the work consistently breaks down.
They know which questions keep returning.
They know which information never arrives when it’s needed.
They know which parts of the process exist only because everyone has quietly learned to work around them.
Leadership still decides what matters most.
Technology still determines what is possible.
But understanding the work creates a much stronger foundation for both decisions.
A different way to begin
This has become one of the questions I find myself asking most often.
Not, Where can we use AI?
But,
Where is the work asking for something different?
Sometimes the answer will be AI.
Sometimes it won’t.
Sometimes the real opportunity is a decision that happens too late.
A handoff that creates unnecessary friction.
A policy that no longer reflects how the organization actually operates.
Or a process that was designed for yesterday’s business rather than today’s.
Understanding the work doesn’t guarantee AI is the answer.
It simply makes it much easier to recognize when it is.
Organizations are asking AI what it can do before asking their people what needs to change.
Reversing that order won’t slow innovation.
It may be what allows organizations to create more meaningful value from it.
Start with the work. Then decide what technology can help it become.