The best automation targets never make the CEO's list

When a CEO brings me in for an AI readiness audit, the first conversation is always about what they want automated. It's a reasonable list. Reporting, proposals, maybe something with scheduling. The things you can see from the corner office.

Then I go sit inside the building and watch people work, and the real list looks nothing like it.

Last week I was watching the finance lead at a contractor I'm auditing walk me through her invoice process. I asked her, "How do you get the invoices into the folder?" She said they come into her email, and she saves them and does it kind of whenever they come in.

For me, I was like, "Bingo. We're going to grab those from your email. We're going to drop them in a folder instead."

Her reaction was, "Oh yeah, I guess this could be done."

She wasn't hiding that task from me. She just doesn't think about it. She does it really quickly, it takes two seconds, and she's been doing it so long it stopped registering as work.

Why this stuff never shows up

The tasks leadership flags for automation are the visible ones, the ones that get complained about in meetings. The tasks that actually eat the company's time are buried in the habits of the people who are good at their jobs. Speed hides the waste. Two seconds here, a minute there, repeated all day by someone who's fast at it, and nobody ever says it out loud because it doesn't feel like a problem.

You can't get that list from an interview with the CEO. I'm running surveys right now at a 50-person GC, and the information I got from the top and the information I got from the interns were vastly different. Same company, and you'd think they were describing two different ones. You have employees all over the place. You have people that are not using AI at all, and you have people that are using it at night and on the weekends in their own free time. Getting everybody to that same level matters, and you can't do it if you only know what leadership knows.

The field is the wrong first move

Here's where I'll disagree with most of the construction tech world. Every vendor wants to start in the field. Field capture, field apps, data entry at the point of work.

But if your field already refused to enter data into an app once, what makes us think it's going to happen this time because the software has AI in it? The technology was never the problem. The behavior was, and people don't change because the tool got smarter.

It goes back to basic human psychology. We like to be good at things. When you have done something so many times for so long, and you do it well and quick, and you're recognized for doing it well and quick, it's hard to jump into something new that you have to be slow at and not good at and kind of work through those frustrations. You don't want to do that. You just want to go back to what feels good. We don't like discomfort as humans.

Start your AI rollout with your most resistant audience and you burn your credibility on round one. There's a better opening move.

Start where the data already exists

The rule I use is simple. If it exists digitally somewhere and a person is retyping it somewhere else, that's your first automation.

Re-keying information from one program to another. Moving numbers from an Excel into another Excel. Saving attachments out of email into folders. Copying the same project data into a second system because the two don't talk to each other. That's the back office, and it's where the time savings are right now. If your team is still living in spreadsheets to hold it all together, that's usually a sign the company has outgrown them.

Nobody has to change how they work. Nobody has to adopt anything. The information already exists, it's just being moved by hand, and the person doing the moving gets hours back without being asked to become a different person.

Earning your way to the field

None of this means the field is off the table forever. It means you get there in the right order.

The resistant people come around by watching AI take real work off their coworkers first. And when it's their turn, you don't hand them a mandate. You give them more education and more training. Give them some special extra stuff, extra cheat sheets, be there for additional questions. More support, so the uncomfortable stage where they're slow and not good at it yet is as short as possible.

Adoption follows proof. It doesn't lead it.

Where to start

You can't find your real automation list from the org chart, and your best targets are probably invisible to you right now. That's the whole reason my readiness audits happen inside the building, sitting next to people while they work. If you want to see roughly where your team stands, the AI readiness quiz is a two-minute start, and the systems I build get wrapped around how your company actually runs instead of forcing you to bend to a platform. If you want to talk through where your company's list actually lives, book a free 30-minute audit and I'll tell you honestly where I'd start.

---

Kelly Stephens is the CEO of EyeOn Automations, a custom software and AI company for construction and real estate operators. She spent 14 years in construction operations at PulteGroup before building software for the industry. Based in Southern California.