The 13 best AI podcasts for product managers, ranked
The best AI podcasts for product managers: how teams build products on LLMs, run evals, set AI strategy and use AI in their own work. Product-side shows, not research papers or news hype. Ranked for what a PM can apply.
We did not rank these from memory. TL;WL has transcribed 377 episodes of 2 of these shows to build its research library, so we know what each one actually covers.
Claire Vo, founder of ChatPRD (an AI tool for writing product docs) and former Chief Product Officer at LaunchDarkly. The show is part of the Lenny's Podcast network.
screen-share interviews where a guest shows one AI workflow, plus short host-only tool reviews · 25-45 min · 1-2 episodes a week · since 2025
Why it is #1: It is the most directly usable AI show for PMs: each episode is one real workflow shown on screen, often by a working PM or product leader.
What it is like: A guest shares their screen and walks through exactly how they use a specific AI tool for real work, step by step. Short and concrete; you can copy most workflows the same day.
How do I use Claude Code as a PM if I don't write code?
How can I turn customer interviews into a first-draft PRD with AI?
How do I build a prototype with AI before I write a spec?
How do I set up evals for an AI feature my team is shipping?
Start with
Evals, error analysis, and better prompts: A systematic approach to improving your AI products | Hamel Husain (ML engineer).A clear walkthrough of how to find and fix failures in an AI product, from one of the best-known evals teachers.
Claude Code for product managers: research, writing, context libraries, custom to-do system, and more | Teresa Torres.Teresa Torres shows how a non-engineer uses Claude Code for research, writing and task management.
Lenny Rachitsky, writer of Lenny's Newsletter and former product lead at Airbnb.
long-form interviews with product and growth leaders · 60-95 min · weekly · since 2022
Why it is #2: Its AI episodes bring in people running AI products at OpenAI, Anthropic, Google and LinkedIn, and Lenny asks for the specific practices, not opinions.
What it is like: Long, structured interviews where Lenny pushes guests for specific advice. Not an AI-only show, but a large share of recent episodes are about building AI products and how AI changes the PM job.
How do I write evals for an AI feature, and who on the team should own them?
Why do most AI products fail after launch?
How should I price an AI product?
What does a PM do differently at an AI lab like OpenAI or Anthropic?
Start with
Why AI evals are the hottest new skill for product builders | Hamel Husain & Shreya Shankar (creators of the #1 eval course).The clearest explanation of evals for product people, from the two people who teach the most popular evals course.
Why most AI products fail: Lessons from 50+ AI deployments at OpenAI, Google, and Amazon.Common failure patterns from 50+ AI deployments, with what to do instead.
Teresa Torres, author of Continuous Discovery Habits and founder of Product Talk.
case-study interviews with the team that built one AI product · 55-80 min · every two weeks · since 2025
Why it is #3: It is the only show built entirely around detailed case studies of shipped AI products, and the host is a respected product discovery teacher who asks product questions, not only technical ones.
What it is like: Teresa talks with the founder, designer and engineers behind one AI feature and goes through how they found the problem, prototyped, and dealt with RAG, agents and evals. Each episode is one product told from start to finish.
How did a real team decide whether to use RAG, agents or a simple prompt?
How do I build evals from the conversations where my AI failed?
How do I move an AI prototype into production without losing quality?
How do product discovery habits change when the product is an AI feature?
Start with
Debugging AI Products: From Data Leakage to Evals with Hamel Husain.Hamel Husain on how to find what is going wrong in an AI product before writing evals.
From Prototype to Production: How Perk Built a Voice AI Agent That Makes 10,000 Calls a Week.A full account of taking a voice agent from prototype to 10,000 calls a week.
Aakash Gupta, writer of the Product Growth newsletter and former VP of Product at Apollo.io.
interviews and tutorials with PMs and product leaders · 55-75 min · about weekly · since 2024
Why it is #4: It covers the AI PM toolset in more episodes than any other PM show, with guests who are working PMs at OpenAI, Google and Meta.
What it is like: Most recent episodes are hands-on AI PM topics: a guest demos Claude Code, Codex or an eval setup, or explains what AI skills PMs are hired for now. Practical, with some career and interview content mixed in.
When should I use chat, Cowork or Claude Code for product work?
What AI skills do hiring managers look for in an AI PM?
How do PMs at OpenAI use Codex day to day?
Start with
How to Build Frontier-Lab Quality Evals with Daniel McKinnon, ex-PM at Meta, Google.A PM-level explanation of how AI labs design evals, and how to copy the approach.
PM's Guide to Claude - When to use Chat vs Cowork vs Code, with Pawel Huryn.A practical guide to which Claude tool to use for which PM task.
Peter Yang, writer of the Behind the Craft newsletter; he led product at Roblox, Reddit, Amazon (Twitch) and Meta.
interviews and live tutorials with product leaders and AI builders · 35-55 min · weekly · since 2024
Why it is #5: It mixes interviews with full tutorials, so you often come away with a method you can follow, and it has kept a steady weekly schedule.
What it is like: Often a guest builds something live, such as an eval, a prototype or an agent, while Peter asks how to repeat it. Some episodes are consumer AI use cases rather than product work.
Hugo Bowne-Anderson, independent AI and data science consultant and educator, formerly at DataCamp, Outerbounds and Coiled.
long-form technical conversations with AI builders · 60-140 min · every 1-3 weeks · since 2022
Why it is #6: For PMs who need to work closely with AI engineers, it explains evals and agent reliability in more depth than any product show.
What it is like: Long, careful conversations about agents, evals and how AI systems behave in production. More technical than a PM show, but the evals episodes are some of the most thorough available.
What is the right way to start evals for an LLM product?
What changed when real users started using a production AI agent?
How do teams build reliable coding agents?
What are common mistakes in AI evals?
Start with
Episode 50: A Field Guide to Rapidly Improving AI Products -- With Hamel Husain.Hamel Husain's method for improving AI products through error analysis.
Episode 60: 10 Things I Hate About AI Evals with Hamel Husain.The common evals mistakes, and what to do instead.
swyx (Shawn Wang), founder and CEO of Latent Space, with co-host Vibhu Sapra. Alessio Fanelli co-founded the show.
interviews with AI engineers and founders, plus short 'Lightning' episodes · 20-110 min · several episodes a week · since 2023
Why it is #7: It is the main show of the AI engineering field, so it is where you hear early how builders are using new models and tools.
What it is like: Fast, technical interviews with the people building AI tools, models and infrastructure. Useful for PMs who want to know what engineers are working with, but much of it is aimed at engineers.
Dan Shipper, CEO and co-founder of Every, with co-host Natalia Quintero. Formerly called AI & I.
interviews about how people use AI in their work, plus behind-the-scenes episodes on Every's own team · 30-80 min · weekly · since 2023
Why it is #8: Dan Shipper runs an AI-first company and asks practical questions about how guests work, so you get real workflows and product thinking in one show.
What it is like: Conversations with founders, writers and operators about how they use AI day to day, often with the tools on screen. Every's own team shares what is working for them as an AI-first company.
Carlos Gonzalez de Villaumbrosia, CEO and founder of Product School.
interviews with chief product officers and product executives · 20-45 min · weekly · since 2017
Why it is #9: It gives the executive view of AI product strategy in short episodes, with a new CPO-level guest most weeks.
What it is like: Short executive interviews, now mostly about how product leaders at companies like Asana, Toast, Deel and Replit are building AI into their products and teams.
How do I add AI agents to a product that was built before AI?
How are other CPOs measuring the return on AI features?
How should a product org change its structure for AI?
How do I turn an AI prototype into production code engineers trust?
Start with
Asana CPO on Why Every Employee Is Now an AI Eval and What That Means for How Enterprises Actually Capture AI ROI | Arnab Bose | E302.How Asana thinks about AI evals and capturing return on AI in enterprises.
Toast VP of Product on Building Agents Into Pre-AI SaaS; Shipping AI Into a 14-Year-Old Platform and Building for the Real Economy.Practical lessons on adding agents to a 14-year-old platform.
Mike Belsito presents it for the Mind the Product team.
solo weekly news recap for product people · 12-18 min · weekly · since 2026
Why it is #10: It is a short weekly AI news show written for product people rather than investors or engineers.
What it is like: A 15-minute recap of the week's AI and product news, framed around what it means for people who build products. Short and easy to keep up with.
Bring it questions like
What does this week's AI news mean for my roadmap?
Why are some companies blocking AI shopping agents?
What can product teams learn from recent AI security incidents?
How is AI changing how companies hire and staff teams?
Start with
How Anthropic’s agents work 6× more than humans.What one AI company's internal data says about agents and team structure.
Product lessons from a month of AI security incidents.Recent AI security incidents turned into lessons for product teams.
Daniel Whitenack, founder and CEO of Prediction Guard, and Chris Benson, AI and autonomy research engineer at Lockheed Martin.
two co-hosts plus a guest · 45-55 min · weekly · since 2018
Why it is #11: It has run weekly since 2018 and explains how AI is actually being deployed, which helps PMs talk to engineering and security teams.
What it is like: Two hosts and a guest talk through how AI is deployed in real organizations: agents, security, open models and enterprise rollouts. Plain explanations, but the angle is more engineering and IT than product.
Jeremy Utley of Stanford's d.school and Henrik Werdelin, co-founder of BarkBox and the venture studio prehype.
two co-hosts interviewing a business leader · 50-70 min · every two weeks · since 2023
Why it is #12: It is a steady source of real examples of AI adoption from leaders at companies like Logitech, Zapier and OpenAI.
What it is like: Conversations with executives and researchers about how organizations adopt AI and how work changes. Focused on leadership and teams rather than product building.
Sam Charrington, industry analyst and host of TWIML.
long-form interviews with AI researchers and practitioners · 50-70 min · about weekly · since 2016
Why it is #13: It has run since 2016 and its practitioner episodes on agents and evals in large companies are useful background for AI PMs.
What it is like: One-on-one interviews that range from ML research to how companies like Airbnb and Capital One build agent systems. More technical than product-focused.
What goes wrong when teams build agentic products today?
How do I find the agent failures my evals miss?
How do large companies deliver multi-agent systems?
What are the security risks specific to AI agents?
Start with
The Brutal Reality of Building Agentic Products Today with Daniel Jeffries - #782.A frank view of what is hard about shipping agentic products right now.
How to Find the Agent Failures Your Evals Miss with Scott Clark - #767.Practical approaches to testing agents beyond standard evals.
Best for: Technical PMs who want to understand the engineering behind AI products
Skip it if: You want PM methods; many episodes are research-level.
Practical depth first: does an episode leave you with something you can use at work on Monday? Then the quality of the hosts and guests, whether the show is still publishing, and how consistent it is. Fame counts for little. Listener numbers are shown only when the show or a credible source has published them, with a link to where; we never estimate them. Facts were checked in October 2026.
Your question. Every show at once.
Have a burning business question you cannot find in one podcast episode? Describe it. TL;WL researches what founders and operators said about it across these shows and hundreds more, then narrates one private episode for you, with every quote linked to the minute it was said.
Free during the beta: 2 episodes a week of up to 15 minutes.
