For decades, IT departments have had a problem.
Some of their smartest people couldn’t communicate with the business.
These were brilliant people. People who could write thousands of lines of code, solve problems that would make your head explode, and build systems that kept entire companies running.
But ask them to sit down with someone from Sales and figure out what that person actually needed? Uh-oh.
There was even a term for some of these folks. “Flat food people.”
The idea was simple. Put the programmer in a room with a computer, slide pizza under the door, and wait for wonderful software to come out.
(Personally, that never worked for me. I prefer deep-dish pizza. They don’t fit under the door)
To be clear, I’m not saying this described everyone. I’ve worked with plenty of technology professionals who were terrific communicators.
But the problem was real.
Companies spent decades trying to fix it.
Communication training. Business relationship management. Organizational changes. Workshops.
We even created roles to help bridge the gap between IT and the business.
And while those efforts helped, the underlying challenge remained.
Wouldn’t it be better if the people building the systems understood the business themselves?
And now artificial intelligence might finally force us to make that happen.
AI Is Changing the Game
AI is getting good at writing code. Really good.
Case in point: I recently added four new HTML pages to my website, www.philbarth.com. Claude wrote the code for all four.
In under four hours!
And I didn’t have to buy it pizza.
AI can also help test software, find bugs, write documentation, and handle a lot of the tasks that used to keep entry-level programmers busy.
Which raises an interesting question:
What happens to those entry-level jobs?
According to a Stanford University study updated in August 2026, employment among software developers ages 22–25 dropped approximately 19% from late 2022 to June 2026.
Meanwhile, employment among several older age groups increased.
Now, before we grab our pitchforks and chase AI out of town, the study doesn’t prove AI caused that decline. There are other possible explanations.
Maybe AI isn’t the reason companies aren’t hiring as many entry-level programmers.
Maybe AI is the reason they can afford not to.
Imagine you’re an IT manager.
Your headcount is frozen. You’ve got experienced programmers who know your systems inside and out. And you’ve got enough work to keep everybody busy for the foreseeable future.
Along comes AI.
Now maybe your experienced programmers can accomplish more without adding people.
Your budget will look better. Your projects will keep moving. Your boss will be happy. Life is good.
Until you consider one little problem.
Where Do Tomorrow’s Experts Come From?
Experienced IT people have something that’s incredibly valuable.
They know the stuff that isn’t written down anywhere.
Why does that old program work the way it does?
Why does changing one little field in a database cause everyone in Finance to lose their minds?
The experienced people know the answer because they’ve spent years dealing with the consequences.
AI can help capture some of that knowledge. But somebody still needs to understand it. And know when something doesn’t look right.
Here’s the real problem:
If companies stop hiring inexperienced programmers, where do the experienced programmers of tomorrow come from?
You can’t hire someone with 20 years of experience if nobody hired them 20 years ago.
That’s a problem we can’t ignore.
Maybe We’re Asking the Wrong Question
For decades, we’ve hired IT professionals because they were good at technology. We put them to work building systems and hoped they’d eventually learn enough about the business to understand what they were building and why.
Sometimes it worked, sometimes it didn’t.
Now AI is changing the work that entry-level programmers have traditionally done.
And that means companies may have to change the way they develop their future IT professionals.
Perhaps it’s time to stop assuming that the best place for a new programmer to start is writing code.
Maybe AI is going to force us to solve a problem we’ve been struggling with for decades.
Coming Next: What If We’ve Been Training IT People Backward?
What if the best way to develop a great programmer isn’t to start them programming?
In part two, I’ll share an idea for developing IT professionals who understand the business before they start building systems for it.
We’ll also tackle a question that young people considering a career in technology might be asking:
“I love computers, but AI is taking over programming. Should I still go into IT, or should I learn which end of a wrench to hold?”
And there’s another group whose importance I believe AI is going to increase dramatically.
Business Analysts.
That’s the subject of Part Three: AI Will Make Business Analysts More Important Than Ever.
Your turn: Do you think AI will finally help solve the communication gap between IT and the business? Or are we creating an even bigger problem by reducing entry-level opportunities?
Research
Stanford Digital Economy Lab (August 2026) — Canaries in the Coal Mine? Six Facts about the Recent Employment Effects of Artificial Intelligence.
The approximately 19% decline in software developer employment among ages 22–25 is an estimated reading from Figure B.3, page 62, and is not a measure of job losses directly caused by AI.
https://digitaleconomy.stanford.edu/app/uploads/2026/08/Canaries_August2026.pdf#page=62
About Phil
Phil Barth is a keynote speaker, author, and recovering corporate IT professional who spent much of his career at International Paper. Today, he helps organizations understand and use artificial intelligence while keeping people at the center of the conversation. His presentations combine practical insights, plenty of humor, and rubber chickens. Lots of rubber chickens.
Learn more at www.philbarth.com.