Six Steps for Landing Your First Product Role in an AI World
How to build real, job-ready experience before you get your first role
What’s the Problem? is publication exploring the future of product building, written by a longtime product leader whose experience spans both the largest (Amazon, Airbnb, CBS) and the smallest (Haven, SET, Jacent) companies around. Like any good product discussion, we start with the question: What problem are we trying to solve?
What’s the problem? What should you do to break into product (or any corporate entry-level job) in a world reshaped by AI?
Last week I wrote about how entry-level roles are changing and why companies shouldn’t eliminate them, but redesign them. This week I’m flipping the script with six concrete steps you can start taking today to ready yourself for your first entry-level role in an AI world.
The old paradox has not changed: you need experience to get a job, but you need a job to get experience. What has changed is the bar. AI has eliminated most of the operational work entry-level workers used to start with, so employers expect you to already know how to use AI effectively and have real, hands-on experience — not just coursework.
Here’s the good news: it’s actually easier now, not harder. One motivated person can run an entire product process with their AI buddy. The advice I’ve given to countless mentees still applies, but AI supercharges it: if you need experience, you can create it.
This post walks through the six things I’d do today if I were entering the workforce. I use product as the example, but the approach works across many entry-level knowledge jobs. (This post isn’t a “how to do product” walkthrough — there are plenty of those — but I’ll link to helpful resources for each step.)
1. Pre-Game: Think about your current work like a problem to be solved.
Before you build anything real, you need to shift how you do your school or every day work. Step 1 is about developing repeatable, AI-assisted workflows — the same muscles you’ll use in every later step. If you can learn to treat your everyday tasks like small, reusable projects, you are well on your way to mastery of AI orchestration.
If you’re in school, you already have the perfect training ground. Take a typical assignment — say, analyzing a company’s strategic position.
If you aren’t in school, treat your job hunt as the project. Record your mock interviews, analyze job descriptions, and iterate on your resume using the same loop.
For students, the old workflow looked like: Google → library (is that still a thing?) → notes → outline → draft.
What you are doing is building an AI-native workflow that’s faster, more rigorous, and reusable across everything you touch.
The new workflow is a loop:
Capture inputs: Record the lecture (with permission), run an AI summary, and upload your notes into a single project workspace/Agent.
Generate a starting point: Ask your AI to synthesize the key research questions, surface relevant sources, and outline initial angles.
Develop your thinking: Read the sources it finds, debate your argument with the AI, iterate until the logic holds.
Draft → critique → refine: You write the draft (yourself). Then run it through your AI for structural feedback, gaps in reasoning, and a clarity pass.
This isn’t about getting AI to do your paper for you, it only works when you still think, write, and decide. What you are doing is building an AI-native workflow that’s faster, more rigorous, and reusable across everything you touch. And yes: this is how people are already working in most knowledge jobs.
2. Solve A Real Problem
Step 1 was about building AI-assisted habits using the work you already have. Step 2 is where you create real experience you can actually talk about in interviews.
You are going to pick a real problem and build a real solution. It doesn’t need to be big; it just needs to be yours. One solid project is enough to anchor your story in interviews.
Look for something you or your friends struggle with, or help a nonprofit or early-stage startup that needs support. Real users instantly make your work credible. This could be a personal to-do list app, a party tracking app, a simple word game—simple ideas are totally fine as long as they solve a real need.
Create a dedicated project space (a GPT/agent that stores files and context) and add a brief doc describing the problem you’re trying to solve and the context behind it. This becomes the foundation for your next steps.
Resources:
Assessing Product Opportunities - Silicon Valley Product Group (Marty Cagan)
How to Get Startup Ideas (Paul Graham)
3. Set a goal
You’ve identified a problem worth solving, but how will you know you solved it? Using one of the many methods for goal setting, figure out at least your north star goal, and ideally a few input metrics or leading indicators and discuss them with your AI (Bonus tip: ask how using this metric could go wrong).
Pick the one that makes the most sense, and ask your AI to draft a quick goal doc. Store that in your project workspace so everything stays connected.
Resources
Choosing Your North Star Metric | Future (Lenny Rachitsky)
4. Create A Strategy
Now that you know the problem and the goal, you’re ready to ideate! Ideally grab some friends, or family, or brainstorm on your own, then uplevel it with AI. Record your thoughts (or the group discussion), upload the transcript to your project, and ask your AI to extract and clean up all the ideas.
Put these ideas into a simple spreadsheet for prioritization. Start with an easy framework like impact vs. effort. Fill it out, upload it back to your AI, and have a discussion about what to adjust. Iterate until you land on a clear roadmap.
Resources:
How Product Managers Use Product Ideation to Build Better Products
RICE: Simple prioritization for product managers (Sean McBride)
The Ultimate Guide to Product Management Prioritization Frameworks
Impact Effort Matrix & How to Use One + Examples (Carlos Gonzalez de Villaumbrosia)
5. Build It!
Use your prioritized ideas to define your MVP (have you read Lean Startup yet? Please do). Pick any requirements format you like. I prefer dictating rough thoughts to my AI and letting it ask clarifying questions, flag edge cases, and clean everything into a structured requirements doc. Save that to your project. As a bonus, ask it to create you a design style guide for your project by interviewing you.
Next, build it (see technical and non-technical builders below). This is the fun part, you get your first version out, and I can 99.9% guarantee it’s not gonna be quite right. Walk through it while recording your reactions, note what’s off, and feed that transcript back into your AI to revise. Revise, rinse and repeat until you get what you want.
Resources
User Stories and User Story Examples by Mike Cohn - Mountain Goat Software
Lovable (Good for non-tech)
Bolt AI (Good for non-tech)
6. Test, measure and iterate
Hopefully you built something that is useful to someone other than you, because it is time to get it in front of real people. Run quick user tests, record sessions (with permission), and collect any notes, messages, or metrics that come in. Upload everything into your project.
Create a prompt that asks your AI to pull out common themes, pain points, and opportunities. Add those insights to your roadmap and re-rank them against your existing priorities. This is where you’re building your continuous discovery/development process. Prioritize your new finding against your already existing roadmap, and iterate.
Resources
The Build-Measure-Learn Feedback Loop (Eric Ries)
How to Use the Build, Measure, Learn Loop In The Product Development Process
The Next Era of Product: Continuous Everything (Justin Eckhouse)
Bonus: Growth
Cherry on Top” for interviews. “Most entry-level candidates stop at building. If you can walk into an interview with actual data on how you acquired your first 10 customers, you are instantly in the top tier of candidates. Because your AI agent already understands your product and market, it’s a great partner for building a simple growth strategy. Ask it to suggest relevant growth tactics, then have it draft ad copy, design basic visuals, or create social posts to test.
You don’t need a full growth engine—just a few small experiments are enough to show you understand how products find users. Though the sky is really the limit here on the amount of growth tests and experimentations you can do.
Resources
How to Achieve a Product-Led Growth Strategy (Ally Heinrich)
Growth Product Manager Skills Every PM Needs to Succeed (Shahzad Shaikh)
Breaking into product in an AI-accelerated world isn’t the apocalypse; it’s just a new landscape with a different set of rules. The tools have changed, the expectations have shifted, and the bar has moved — but you are absolutely capable of meeting it. Every step in this post is learnable, repeatable, and fully within reach for anyone willing to put in the time. You CAN do this! Start small, build something real, iterate like crazy, and you’ll be far ahead of where most entry-level candidates ever get.
Where’d I get it wrong? I’d love to hear it, please leave a comment (or have your bot do it) and subscribe.