How Generative AI Changed My Career

This change has come and gone, to some extent. And we can now see what's filling the upper shelf.

Generative AI is ubiquitous now, and that ubiquity has split people, again, into two camps. You may recognize the shape of this debate. It echoes the one I described two years ago between the calculator gatekeepers and the calculator believers.

Camp one is producing what some are calling AI slop: endless fake videos, AI-generated memes, deepfakes of celebrities, images of things that never happened, content that piles up in feeds like digital litter, volume without value. The tools lowered the floor so far that anyone can generate anything, and plenty have taken that as an invitation. A lot of it is fun, some problematic.

Camp two is pressing in on real use cases in professional settings. These are the people who have quietly gone to work, drafting reports, summarizing transcripts, isolating themes from open-ended survey responses, extracting signals from documents that would have taken days to read. They are not showing off. They are solving problems, and they are getting faster and better at it every day.

This distinction matters because it shapes how we think about AI's actual impact. The slop is public and persistent. The professional use is quiet and has great potential. But the quiet camp is where the real transformation is happening, and I'd argue it's where the upper shelves are being built.

There's another narrative running alongside all of this, and it is hard to miss: AI is coming for your job.

You've heard it. Probably in a headline. Possibly in a staff meeting. The message is that the tools are so capable now that they will simply replace the humans who used to do the work. It is a reasonable thing to worry about, and for some roles in some industries, the moment is probably coming or has already arrived.

But for me, AI did not take my job. It gave me a new one that didn't exist before.

Generative AI is part of marketing packages now, for products, software, services, and of course for its own provider accounts. The shelf metaphor had its limits, it turns out. We've learned what kinds of things it has made available to people who couldn't previously access them: writing (we figured that), but now art, music, coding, graphic design, voice conversation with thought partners, video production, translation, data analysis, language practice, tutoring, and on and on.

And it keeps getting better at all of these, and that rate of improvement is getting faster.

Allow me to share a case study from my own experience, a case where generative AI granted me access to something I previously could not have reached, and wouldn't have even thought of trying.

In August 2023, just before writing my first article on generative AI, I was still a skeptic. My Head of School asked if I'd like to join the AI team and travel to a conference in Manila, learn about its potential, and then guide the school into whatever adoption was going to look like on our campus. I declined, saying, "It's gimmicky. It's just not there yet."

Some time later, I had an encounter with an early adopter who showed me what he'd been doing with it, and it blew my mind. It wasn't gimmicky at all. My explorations had simply been at the shallowest level, I had been in camp one without realizing it. I immediately went back to the Head of School to see if there were still tickets to the conference. There were.

The conference passed in a hurry, something like drinking from a fire hydrant. I tried my best to at least come away with names of software to explore later. AI companies were popping up like mushrooms, and software companies were racing to get a bot in their platform first. I came back and explored when I could, but I had a job to get back to. As Deputy Principal of Secondary School, my AI use was limited by capacity. My early message to the school was, "Here is generative AI. Here's its definition. Here’s a cool picture of Darth Vader sitting on a countryside porch!," and, as everyone else was doing, "This slide show was made by AI!" Abracadabra! Everyone oohed and ahhed. I'd rolled out AI to our school. Check. But I had other things to get back to.

Not long after that, I wrote my calculator article. Then in a weekly meeting with my Head of School, he asked me a very different kind of question than I'd ever received. Our school's leave management system was frozen in time. "Do you think you could build a replacement for it?"

Without a moment’s hesitation I responded, "Yes."

I had no background in software development, no training in application design, no coursework in code, no portfolio of previous builds, and no obvious reason to say yes. What I had was a year of watching what people were saying AI was becoming capable of, article after article, across what felt like every news source. It was obvious to me that all I needed to do was learn how to use it properly. It was a risky move, but I was completely confident it was doable.

The Head of School made time in my schedule and got me a second office. It was only a matter of a day before I understood what it really meant to code something (this was in the days before AI coding platforms that create ready-made apps). By the next day I had a skeleton of a new leave management system. I was on Generative AI just asking what to do, where to put it, how to save it, how to deploy it and how to make it so others could see it. I worked entirely in Google Workspace to produce our school’s first app. It was completely unbelievable to me. My risky commitment was entirely within my grasp. Three days later I presented a first draft to the Head of School as a concept. He immediately gave me the go-ahead to spend all my newly allocated time building the full system.

What an exciting time, I thought. I had suddenly become a software developer.

Almost immediately, the Head of School began making adjustments to accommodate more building time, and soon after, posed a question that changed everything. "If I created a Systems Management position, would you take it?" Without hesitation, I said yes. This series of conversations set my career on a new path, not because AI replaced me, but because AI made me capable of something I had never been capable of before. And an institution noticed.

The role of Director of Systems Management at our school did not exist before the work began. In a standalone international school there are no district offices, no superintendent's office, no centralized administrative infrastructure. The kind of stability those systems offer in national school systems does not exist in international schools. Schools have to make those up themselves. The need to think architecturally about institutional systems has always been present. In our case, it had simply never been prioritized, because filling it had always required resources and expertise that were difficult to assemble. The resource that was most scarce was time.

For three years prior, in weekly meetings with my Head of School, I had been identifying gaps in our school's systems architecture, and then putting them on the back burner. As the newly established Director of Systems Management, now was the time to address all of them.

The Systems Management portfolio now runs to well over a hundred custom-built applications: portals, dashboards, sites, and productivity apps. We now have bespoke teacher evaluation software and human resource management infrastructure, admissions dashboards and administrator-facing tools for analyzing standardized testing datasets and gradebooks, and rapid communication applications that allow teachers and administrative assistants to reach the right stakeholders in a fraction of the time that standard email interfaces require. We have air quality and heat index reporting tools the school nurse can use to push info to staff and families in real time. Every division has a single sign-on portal, each with tiered access levels calibrated to the people who use them. Most recently we are creating a parent-facing portal designed to feel as intuitive as a phone screen, with every resource a family might need gathered into a single place. And the list goes on and on.

The driving philosophy behind all of it is simple, and it has not changed since the first system went live: a one-stop shop with the fewest clicks possible. Providing a single access point where people can find everything, and then use the fewest clicks possible to access what they need, is a gift of time returned to the person on the other end.

None of this existed before the new AI tools made it possible for someone like me to build it. I am not a developer with a programming background who works alongside educators to fill their needs, an outsider pressing in. I am an educator with14 years of experience working with youth and families in the classroom and in the office, who has learned how to develop software that meets needs from a perspective inside the world of education. An insider, reaching out to bring software development in.

So, for me, this is what became accessible on the upper shelves: software development.

If I predicted AI would be like a word calculator to help people access writing, a very limited concept which would pale in comparison to what would truly become available, then for me, AI has provided a coding calculator. Access to a process that allows me to take any problem any stakeholder can present, and collaborate directly with them as my client to create a solution beyond what they could have hoped to ask for.

AI in our context has opened up a whole new line of inquiry that was never possible before. What if we could provide a solution for any friction point in our school's daily operations? What does it mean to have someone who understands the workflow of a school, the pressure points in the faculty's daily experience, the communication breakdowns that quietly drain energy from an organization, someone who can translate that human experience into the architecture of the systems they build? 

These are the questions I have spent the last two years answering in practice.

If systems architecture is now on the lower shelf, if an educator with no background in software development can say yes to a software development project, make it work, sustain it, and build a career around it, then what occupies the upper shelves in its place now? What fills the space formerly held by the exclusive domain of engineers, and what will it mean for the institutions and the people who begin to reach for them?

We don't have a complete answer yet. But here's what I notice: the people who are finding the answer are in camp two. They are not producing slop. They are pressing in on real problems, in real professional contexts, and they are building real solutions, and finding roles, and doing work that didn't have a name two years ago.

As for the camp-one fear that AI will take your job, I'd suggest a different framing. The question isn't whether AI will take your job. The question is whether you are paying close enough attention to what the tools make newly possible, and whether you are willing to say yes before you know exactly how you'll deliver.

I have a hunch that more people will find themselves reaching that upper shelf very soon—whatever that looks like for them. And I believe that in years to come, we’ll look back and recognize how significant this moment of change was, and that we were there to witness it as it unfolded.