Engineers were already powerful. AI makes them more so. That’s the expected story. Here’s the one nobody is telling.
Over the past year, I’ve been teaching the operators and founders on my own team to build with AI. Not engineers — marketers, project managers, operations leads. What we observed was counterintuitive enough that we built a program around it.
Engineers who adopt AI get better at what they already do. That's real, and it’s significant. But the transformation we kept watching happen with non-technical professionals was something else entirely. People who had never built anything were suddenly making things from scratch. People who used to wait weeks for an engineer to solve a problem were solving it themselves by Thursday. The gap between what they could do before and after AI was not incremental. It was categorical.
That gap — not any comparison to engineers — is the story worth paying attention to.
Who Gets the Most Out of AI?
While AI helps engineers work faster, its most transformative impact is on non-technical domain experts. By using rich, precise domain knowledge instead of coding skills, these experts can use plain English to build tools, automate workflows and solve problems independently, eliminating their reliance on a technical middleman.
AI Gives Engineers the Boost We Expected
Engineers are natural AI adopters. They’re already comfortable with using digital tools, iterating and working within systems. Add AI to that foundation, and they get faster, they automate more, they ship more. The compounding benefits are real.
But engineers were already operating at a high technical baseline. The boost is meaningful, but it’s drawing on their existing skills. They understand the problems they’re solving, and they had ways to solve them before. AI just makes those ways faster and cheaper.
What engineers bring to AI is technical fluency. What they sometimes lack is the thing that makes AI outputs actually useful: Deep knowledge of the domain the output is supposed to serve. An engineer can build a tool for healthcare operations faster than ever before. But they still need a healthcare operations director to tell them what the tool should actually do, where the real bottlenecks are and what a good output looks like.
AI Gives Non-Technical Experts a Bigger Advantage
The more interesting transformation is happening with the people who know their domains deeply but previously had no way to build anything to serve them.
A healthcare operations director who has spent 15 years dealing with prior authorization workflows, staffing models and compliance requirements — she understands the problems in her field better than any engineer ever will. For most of her career, turning that understanding into a functional product required a technical intermediary. She’d describe the problem, hand it to an engineer, wait, review a piece of software that missed half the context and then iterate. The cycle was slow and the output was always an approximation of what she actually needed.
AI collapses that gap. She can now describe her problem in plain English, with all the nuance and domain context she carries, and get back a product that reflects it. She’s not replacing the engineer. Instead, she’s removing herself as a bottleneck. She solves the first layer of the problem herself, faster, and the work that reaches an engineer is better scoped and further along in the development cycle.
My co-founder Ravi is also CEO of another company. He’s not an engineer. A few years ago, he faced entire categories of problems he simply could not address without hiring technical experts. Today, he builds tools, automates workflows and has taught his team to do the same. That’s not because he learned to code, but because AI gave him a way to apply the deep business knowledge he already had. The constraint wasn’t intelligence or understanding. It was access.
That’s the pattern we keep seeing. Domain experts don’t need AI to make them smarter about their field. They already are. They need AI to remove their dependency on a technical middleman that was never really about the ideas. It was about the implementation.
Domain Depth Is the Real Competitive Advantage
The variable that separates people who get transformative results from AI and people who get marginal ones isn’t technical fluency. It’s domain depth.
When you know your field precisely — when you can describe not just what you want but why it matters, what the exceptions are, what a bad answer looks like — AI can do something with that. Vague inputs produce vague outputs. Rich domain knowledge produces rich outputs.
Non-technical professionals have spent years or decades building that depth. AI has finally given them a tool that can work with it directly, without translation.
This is why a marketing coordinator who becomes a serious AI user is, in our experience, genuinely unstoppable. She already knows the audience, the message, the constraints, the brand voice, the competitive landscape, the history of what has and hasn’t worked. Add AI to that and she can now produce, test and iterate at a speed and volume that would have required a team of people two years ago. She didn’t need to learn marketing. She needed access to a tool that could keep up with what she already knew.
How Should Organizations Think About AI?
Most companies have built their AI strategy around their technical teams. They assume the engineers will figure it out and then teach the rest of the organization. This makes intuitive sense but gets the order wrong.
The people who understand where the real problems are — the inefficiencies, the manual work, the decisions that take too long — are almost never in the engineering org. They’re in operations, in sales, in client services, in marketing. They’re the ones who know which workflow that should take 20 minutes actually takes four hours because they’re the ones doing it.
Give those people access to AI tools and enough structure to get started, and they’ll identify the highest-value applications faster than any centralized technical team will. Engineers still do the work that requires engineering. But the range of tasks domain experts can handle themselves expands significantly, all of which changes what the engineering team’s time is actually spent on.
We run MakerSquare without any engineers on the team. That’s not because we're anti-engineering, but because the work of building a training program — curriculum design, operations, marketing, client relationships — is domain work. AI handles an enormous amount of the implementation layer. What remains requires judgment about our field, our students and our business. That’s exactly what we already had.
The One Thing Non-Technical Professionals Need
The transition isn’t automatic. There’s a real learning curve. It’s just not the one most people expect.
Non-technical professionals don’t need to understand how AI works. They need to learn how to describe their problems with precision. How to give context that shapes the output. How to evaluate a response critically and iterate toward something better. They need to move from “give me an answer” to “here’s my situation, here’s what I know, here’s what a good answer actually looks like.”
Instead of typing “write me a marketing email,” a coordinator who knows her audience types: “I’m writing to healthcare operations directors who are overwhelmed by manual scheduling. They’ve heard AI promises before and are skeptical. Write an email that leads with the problem, not the solution, and ends with a single low-friction request.” The first prompt gets a generic email. The second gets something she might actually send.
That shift takes most people a few hours of deliberate practice. Once it clicks, however, the domain depth they spent years building becomes the primary input. Then, suddenly, the thing that made them good at their job also makes them fast, autonomous and far harder to compete with.
Engineers were already in the game. Non-technical domain experts are just getting started. And the gap they’re closing is bigger than most people realize.
