How AI Is Evolving PM Skill Sets

AI job postings as a percentage of all positions nearly doubled between 2019 and 2024, representing hundreds of thousands of new roles requiring AI competencies. As of late 2023, there were over 14,000 open AI product manager jobs globally, with nearly 7,000 in the U.S. alone.

We’re past the question of whether AI will replace product managers. The more urgent, practical inquiry is: How will PMs use AI to be more effective, and what does that mean for their skills?

The role is changing fast. We’re watching it shift from coordination and tactical execution toward something more strategic, more technical, and more ethically complex. Success today isn’t about replacing old skills with new ones, but blending timeless product instincts with technical fluency and AI-native thinking. 

The Great Debate: Is the Non-Technical PM Obsolete?

The rapid integration of AI into product development has ignited a widespread, and often anxious, debate: Is the era of the non-technical PM over?

Adam Judelson doesn’t mince words. “The non-technical PM is dead,” he said in a recent Productboard webinar. In his view, the middle layers of product management—focused solely on stakeholder management or process orchestration—are being automated out of relevance. “If you actually got your employee to use a really good prompt that represents a manager, you probably wouldn’t need most of the middle management.”

At the same time, PMs don’t need to be engineers. The truly endangered species isn’t the PM who can’t write code, but the one who can’t speak credibly about how things are built. The baseline for technical fluency has been permanently raised.

Interestingly, our recent CPO Survey indicates that many leaders continue to emphasize strategic skills in their hiring—yet PMs are also expected to demonstrate AI fluency. The sweet spot lies in knowing enough to shape technical opportunities, without getting lost in the weeds.

What Are the Skills? 

The role of the PM has always been that of an “expert generalist,” a central hub connecting various specialists. As technology becomes more integral to all products, the scope of that “general” knowledge must expand. 

Of course, it’s all about balance. An overly technical perspective can be a detriment, skewing a PM’s ability to view the product from the non-technical user’s perspective. Instead, the new requirement is for a deeper conceptual understanding of the systems they manage.

Strategy Takes Center Stage 

AI can now handle much of the coordination and documentation that used to consume a PM’s time. That means the highest-leverage PMs will increasingly be the ones driving strategic vision, not just shipping roadmap items.

As Adam puts it, the real value today comes from two ends of the spectrum: the “super-IC PMs” who build in the weeds, and the chief product officers who set strategy and kill bad bets. In between? According to Adam, “Managing those people, asking them questions about what they’re doing—I’m just not sure that that is a viable place to hang out anymore.”

Basic Technical Skill Is the Foundation 

Technical fluency isn’t about writing perfect code. It’s about understanding systems well enough to shape ideas, spot feasibility issues early, negotiate for resources, and earn the respect of the development team. The bar isn’t CS-degree high. But you need to know enough to ask smart questions, pressure-test decisions, identify dependencies, and weigh trade-offs with credibility.

How to Develop Technical Fluency

For those worried that they aren’t a technical PM? Start small. “Build something dumb,” Adam advised. “The best thing that you can possibly do is build something really simplistic yourself.”

Find a problem, use AI tools to create a prototype, and follow the thread until you run into something you don’t understand. Then, go learn that. “I try to dedicate some time each week to generative AI learning and trying new tools,” he added.

He suggests deploying simple apps as a crash course. “You’re going to see pretty much every piece of the technical pie except for that heavy DevOps scaling phase.” This kind of hands-on learning builds conceptual understanding—the kind that helps you lead, not just follow.

Safety. Security. Ethics. 

As AI becomes more embedded into products, PMs must help their teams balance innovation with responsibility. This means partnering deeply with security, even if it means wading into technical territory. According to Adam, “This is another plug for why you need to be a little bit more of a technical PM. Usually the things that security wants are pretty DevOps-y in infrastructure. So you've got to build that relationship with security.” 

It also means engaging in ethical questions—about data privacy, model transparency, and unintended consequences. These aren't future problems. They're current responsibilities.

Your 3-Pillar PM AI Skill Stack

Unsure if you’re a technical PM? No problem. Take a look at the skill stack below to assess how many of the below competencies you can check off.

Pillar 1: Foundational AI & Data Literacy

This is the new technical baseline. The non-negotiable knowledge every PM needs to work credibly with technical teams.

• Understanding ML Concepts: PMs don’t need to code, but they must know the basics—how models are trained, the difference between supervised vs. unsupervised learning, and trade-offs like accuracy vs. bias.
• Data Fluency: Data is the raw material of AI. PMs should understand pipelines (collection, cleaning, labeling) and be able to partner with analysts to interpret results and guide improvements.
• Domain-Specific AI Knowledge: Beyond general AI, PMs should know the core technologies in their space (e.g., natural language processing (NLP) for chatbots, computer vision for image products) to spot opportunities and constraints.

Pillar 2: The Ascendance of “Power Skills”

As AI takes on routine tasks, uniquely human skills drive PM value.

• Strategic Storytelling & Communication: Translate complex AI into simple, compelling narratives for engineers, execs, and customers.
• Ethical Reasoning & Responsible AI: Proactively address bias, privacy, and fairness—building trust and accountability into products.
• Emotional Intelligence & Empathy: Champion the user, understand real pain points, and build cross-functional trust.
• Critical Thinking & Judgment: Apply context and product sense to AI outputs; machines provide data, PMs provide judgment.

Pillar 3: Practical AI Fluency

These are the hands-on skills that make AI a daily force multiplier.

• Prompt Engineering: Write clear, context-rich prompts to synthesize insights, draft artifacts, and shape user-facing AI interactions.
• AI Tool Integration: Select, evaluate, and embed AI tools into workflows so they actually improve team productivity.

How to Build the Skills 

Developing these skills isn’t about enrolling in one-off classes or checking boxes. It’s about creating a personal learning program—a self-study guide you can actually stick to. That might include:

Essential Newsletters:

• For General AI Trends: To stay on top of the broader AI landscape, newsletters like The Rundown AI (daily summaries of AI news) and The Neuron (deep dives into AI trends and tools) are essential reading.
• For Product Management Craft: To hone core product skills, Lenny’s Newsletter (deep dives on product and growth) and the SVPG Newsletter by Marty Cagan (expert insights on building great products) remain industry standards.

Must-Listen Podcasts:

• For AI Innovation: To understand where the technology is heading, podcasts like No Priors (hosted by investors Elad Gil and Sarah Guo) and Practical AI (focusing on real-world applications) offer invaluable insights.
• For Product Leadership: For timeless advice on the craft of product management, Lenny's Podcast (in-depth interviews with product leaders) and One Knight in Product (conversations with PMs at all career stages) are highly recommended.

Key Courses & Tutorials:

• For Foundational Knowledge: Several universities and companies offer high-quality specializations for PMs, such as Duke University's AI Product Management Specialization and IBM's Generative AI for Product Managers Specialization.
• For Industry Certifications: Organizations like Product School (Artificial Intelligence for Product Certification) and Udacity (AI Product Manager nanodegree) offer structured programs designed to build AI literacy for product professionals.
• For Hands-On Skills: For those who want to get more technical, Andrej Karpathy's YouTube series on building neural networks is a highly recommended resource for gaining a deep, practical understanding of how LLMs work. For no-code development, numerous tutorials are available on YouTube and platform-specific learning centers.

This isn’t about becoming an engineer. It’s about building structured habits that keep your knowledge current and actionable.

Why is this important? Well…

Learning Is the Skill 

In the past, PMs honed skills like writing PRDs or running meetings. Today, the key meta-skill is learning itself. Systems are changing too fast for static checklists. What matters is how quickly you can pick up new patterns, tools, and concepts.

Adam suggests three daily learning loops:

1. Think for yourself.
2. Ask AI to think.
3. Think together with AI.

“Thinking together with AI is a back and forth conversation. You don’t like the answer? Ask something else—provide more context, keep going. I'm doing this constantly throughout the day. All of my product managers at First Principles and engineers at Common Sense are doing this too.”

Here’s a simple, practical strategy you can follow:

• Prototype something. Even something small.
• Run into a problem. Let that problem guide your learning.
• Use AI tools to accelerate. Don’t just prompt—build with them.
• Partner across disciplines. Security. Legal. Data. You don’t need to be the expert, but you need to credibly represent their inputs.
• Zoom in and zoom out. The best PMs blend high-level strategy with tactical execution. If you know how to use it, AI can help you do both.

Want more? Watch the full webinar on demand.