How New Graduates Can Win in an AI World

Every week, someone in my LinkedIn feed declares that entry-level roles are dead and AI has replaced junior employees. I get why the narrative sticks. AI can draft PRDs, analyze data, summarize research, generate wireframes, and ship functional code. When I talk with computer science juniors and seniors preparing for their first product management interviews, I can feel the anxiety.

But here is what I want them to understand: those tasks were never the product. PRDs, backlogs, and status updates were always just tools. The real job has always been to create clarity, exercise judgment, and build products that solve real human problems.

I have led product teams across streaming, e-commerce, search, and mobile. Across every era and every technology wave, the pattern is the same: when a new tool lowers the cost of execution, the people who thrive are the ones who know what to build and why.

For aspiring PMs, this moment is not a contraction. It is an opportunity. The bar is rising, but so is the leverage available to you. Early-career builders who embrace an AI-native mindset have a real advantage. You are not unlearning years of process built for slower cycles. You can design your entire operating system around modern tools from day one.

So stop asking, "How do I compete with someone who has ten years of experience?" Ask instead, "How do I use AI to compress the gap between my potential and my output?"

Here is where I would focus:

Get fluent in AI-assisted prototyping. You do not need to be a full-time engineer, but you should be able to turn an idea into a working artifact using AI coding tools, lightweight backends, APIs, and version control. A functional prototype changes the conversation faster than any slide deck.

Invest in structured problem framing. AI can generate outputs all day long, but it cannot reliably define the right problem. Breaking ambiguity into components, defining success metrics, and articulating trade-offs are the highest-leverage skills you can develop.

Develop real data fluency. Interrogate dashboards, design simple experiments, and interpret results with nuance. Connecting your decisions to measurable impact signals a maturity that goes well beyond tenure.

Build judgment around models. Learn to design prompts intentionally, evaluate outputs systematically, and define guardrails. In an AI-enabled environment, discernment matters far more than speed.

Practice end-to-end ownership. Pick a small problem and carry it all the way through: clarification, build, testing, iteration, reflection. True ownership builds credibility instantly.

Sharpen your communication. Clarity drives decisions and aligns teams. In faster development cycles, confusion compounds quickly. The person who can make complexity legible will always be in demand.

From my vantage point hiring and building product teams, I am not looking for someone who can produce a document in isolation. I am looking for agency. Can this person take ambiguity and turn it into motion? Can they use modern tools to reduce friction? Can they demonstrate judgment when execution is easy but the consequences are complex?

Look at how leading companies are already defining senior roles. A recent "Senior Product Manager, Builder" posting at Amazon calls for building rapid functional prototypes with AI coding tools, driving ideation through automated data analysis, using LLMs for strategy, and building applications with generative AI patterns like RAG. A preferred qualification is a portfolio of functional projects built using AI tools. Your GitHub profile should be your resume.

That is a senior role today. If the tools are accessible, the documentation is public, and the infrastructure is inexpensive, what exactly prevents a motivated new graduate from developing these capabilities before day one?

There is a useful economic concept here: Jevons's Paradox. When technology dramatically increases efficiency, total demand tends to expand rather than contract. Steam engine improvements did not reduce coal consumption; they increased it, because lower cost unlocked more use cases. The same pattern played out with computing, cloud infrastructure, and smartphones. AI is doing the same thing right now. When execution gets cheaper, more ideas become viable, more experiments get run, and more products get launched. The constraint shifts from effort to judgment, from output to ownership.

The strongest early-career candidates I have seen bring proof. They bring working prototypes, clear analyses, and honest reflections on what failed and why. They operate like builders long before anyone gives them the title, and that signal is impossible to ignore.

If you are early in your career, focus on visible evidence of capability. Ship small but real projects. Publish concise reflections on your decisions. Go deep on one AI workflow instead of skimming many. By the time you apply for full-time roles, you should not have to say you want to be a PM. You should be able to demonstrate that you already think, build, and learn like one.

Entry-level roles are not disappearing. They are becoming more technical, more autonomous, and more outcome-oriented. For those willing to build, this is not the collapse of the first rung on the ladder. It is a redesign that favors the hungry over the credentialed.

This article reflects my personal perspectives on product management and AI. It does not represent the official position of my employer or any affiliated organization.