AI product management in 2026 — what PMs need to learn

What the video covers

Alex Zinoviev, Head of Product at Neo Intelligence and co-founder of GenAI Lab (one of Australia’s largest AI product communities), joins the Experts in the Loop podcast to discuss what the AI product manager role actually looks like in 2026. The conversation covers the gap between AI strategy and AI theatre, the specific skills PMs need to acquire, and how data loops create durable competitive advantages.

Who it’s for

Product managers at any level who are trying to figure out which AI skills to invest in and how to distinguish real AI strategy from performative adoption. Also useful for UX and design professionals partnering with AI product teams who want to understand the PM perspective.

Key takeaways

1. Two types of AI PM — and both need new skills. Zinoviev draws a clear distinction between PMs building AI products and PMs using AI to do their existing work better. Two years ago, using AI as a PM tool was optional. Now it is expected, regardless of whether the product itself involves AI. The transition is underway for most PMs, but many are only beginning.

2. The new skill stack is concrete. Rather than vague advice about “learning AI,” Zinoviev names specific capabilities: understanding large language models (what they can and cannot do), evaluation methods for AI outputs, prompt engineering, and AI automation workflows. These are not theoretical — he describes using them in daily work at Neo Intelligence, including a real example of building AI automation for a sales team.

3. AI strategy vs. AI theatre is the real divide. “Put AI on the roadmap” is not a strategy. Zinoviev argues that a year ago, most CEOs told their CPOs to “add AI” without defining what that meant. The result was experimentation that looked productive but produced no measurable outcomes. Real AI strategy requires connecting AI capabilities to specific business problems with clear success metrics.

4. Data is the actual moat. Better products generate better data, which feeds better models, which produce better products. This data loop is what creates competitive advantage — not the choice of model or the quality of individual prompts. Zinoviev encourages PMs to think about where their product’s data flywheel is and how AI can accelerate it rather than treating AI features as standalone additions.

5. Tool-hopping is expensive. There is an opportunity cost to constantly switching between AI tools and platforms. Zinoviev recommends going deep on a single toolset rather than chasing every new release. AI itself can help shorten the learning curve on chosen tools, making depth more accessible than it was even a year ago.

Worth watching if…

You are a PM who has been told to “learn AI” but does not know where to start, or you are trying to build an AI skill development plan for your product team.