Key Takeaways
- Systems thinking, the ability to break a desired outcome into clear, logical steps, is the skill that most determines AI effectiveness, and most job descriptions don't screen for it.
- Coding syntax is becoming less relevant; what's becoming more valuable is systems thinking: understanding how pieces connect and how to give AI agents structured context to work from.
- Vague goals produce vague AI results. The clearer and more logically structured the input, the better the output, every time.
- When hiring for AI fluency, look for people who can define a problem before they touch a tool, not just people who know which tools to use.
Listen: Systems thinking is the AI skill nobody is talking about.
Every job description for an AI-adjacent role looks roughly the same right now. Proficiency in specific tools. Familiarity with particular platforms. A list of models the candidate should have worked with.
What almost none of them ask for is the skill that actually determines whether someone will be effective—the ability to break a desired outcome down into a logical, step-by-step process and communicate it clearly.
That's systems thinking. And it has almost nothing to do with whether someone can write code.
What systems thinking actually is
Here's what I mean. When a Developer sits down to build something, the actual typing, the syntax, the structure, the specific commands, are the least of it. A thousand lines of code takes maybe fifteen minutes to type. The work is everything that comes beforehand. Understanding the goal, identifying the constraints, and mapping out the logical flow of how one step leads to the next.
That thinking (the ability to construct a sequence of clear, logical steps that lead to a complex outcome) is systems thinking. It's what makes someone effective at working with agentic AI. Not because they're writing code, but because that's exactly what AI agents need from the people directing them.
An agent doesn't struggle because the model isn't powerful enough. It struggles when the person using it can't articulate what they actually want. Vague goals produce vague results. The clearer and more logically structured the input, the better the output, every time.
Why coding syntax is becoming less relevant
This is where I'll push back on a common assumption: that the rise of AI makes coding skills more important than ever.
Knowing specific syntax, the precise grammar of a given language, how to structure a for loop, and how to organize code into object classes is becoming less essential, not more. AI can generate that. What AI needs help with is understanding your business problem, defining the logical steps required to solve it, and translating that into something buildable. That's still a human job.
The Developers I see getting the most out of AI tools right now aren't necessarily the ones with the deepest technical knowledge. They're the ones who can step back from the syntax and think in systems, who understand how pieces of a whole connect, what the goal actually is, and how to give an agent enough structured context to run with it.
The skill that's becoming obsolete is memorizing commands. The skill that's becoming more valuable than ever is systems thinking itself—structuring a problem clearly before you ever touch a tool.
What this looks like in practice
When I start a new agentic AI project, I don't open a code editor. I open a blank document and write a clear, concise paragraph describing the end goal. Then I have the agent interview me, asking the questions it needs answered before it starts building anything. That interview process forces me to work through the logic up front: What are we trying to accomplish, what are the constraints, and what does a good output actually look like?
That's product thinking. It's the same discipline a good UX Designer or Product Manager brings to a project, understanding the goal before you touch the tools. The people I've seen struggle most with agentic AI are the ones who want to start building before they've defined the problem. They hand the agent a vague instruction, get a vague result, and conclude that the technology doesn't work.
The technology works. The problem is that they couldn't describe what they wanted clearly enough for it to help them.
The gap in how we're hiring right now
When hiring managers evaluate candidates for roles that involve AI, whether that's a Developer, a Strategist, a Creative Director, or a Marketing Manager, they're often screening for the wrong things.
Tool proficiency is table stakes and changes fast. What doesn't change is whether someone can think through a problem systematically, describe a desired outcome with precision, and stay focused enough to iterate toward it. Those are the people who get real results from AI, and those are the people most job descriptions aren't specifically asking for.
The tell isn't whether someone knows a particular platform. It's whether they can explain, clearly, logically, in their own words, how they would go about solving a problem they've never encountered before. Whether they can break it down. Whether they understand what information the AI would need. Whether they've thought through what a good output looks like before they start prompting.
That thinking tends to show up in backgrounds that might not look obviously “technical” on paper. Product Managers. Creative Directors with systems-oriented thinking. Designers who've worked at the intersection of process and output. These are the profiles worth paying attention to.
Curiosity is the other half
Systems thinking gets you the structure. Curiosity gets you the iteration.
Working with agentic AI looks like this in practice: you define the goal, set up the context, run the agent, and then you evaluate what came back, identify what's wrong, and refine. And then you do it again. The results improve not because the tool got better overnight, but because you kept going.
The people who give up after the first output that isn't quite right are consistently going to underestimate what AI can do. The people who treat the first output as a starting point, who are genuinely curious about why it went the way it did and how to push it further, are the ones who end up building things that actually work.
When you're hiring for AI fluency, you're not looking for someone who already knows the answers. You're looking for someone who won't stop until they find them.
About the author
Dan heads up Aquent's Software-as-a-Service (SaaS) division that produces RoboHead, a leading project management solution for in-house marketing and creative. Dan started as a web developer at the University of Georgia, working on one of the first web-based course registration systems in the U.S. He then worked as a consultant for Lotus and IBM, overseeing the design and architecture of applications for numerous Fortune 500 companies. He joined Aquent as the director of research and development, helping the company transition from client/server to a multi-tiered web architecture. Dan's mission with Aquent Cloud is to produce software that helps in-house marketing and creative teams maximize their productivity and quantify the value they bring to their organizations. Dan holds a BA and MA in English literature from the University of Georgia. Caught up in the exciting possibilities ushered in by the birth of the internet, he was never able to finish that dissertation and emerge as a leading authority on British Romanticism.
Latest.
Smarter workflows start with automation before adding AI.
Leadership & Management, Engineering & Technology, Innovation & Emerging Tech
Who to hire for Agentic AI orchestration.
Career Advice, Consulting & Operations
The AI talent you need isn’t where you’re looking.
Leadership & Management, Talent Acquisition & Recruitment
