Skip to content
Aakash GuptaAakash Gupta

Complete Course: AI Product Management

From prompting through AI PRDs, fine-tuning, RAG, MCP, and AI Agents, today's episode is a complete crash course on how to become an AI PM. Trailer - 00:00 Why AI PMs Are Paid So Much - 1:25 Effective Prompting for AI PMs - 02:39 Ad: Linear - 09:57 Ad: Miro - 10:42 AI PRD Template - 11:54 Fine-Tuning vs RAG - 16:42 Ad: Amplitude - 19:01 Fine-Tuning Demo: Creating a Yoda-Style AI Assistant - 19:52 RAG Implementation: Connecting Documents to AI Chatbots - 30:03 MCP (Machine-Callable Programs): Working with Multiple Tools - 59:00 AI Agents: Creating Advanced Product Research Assistants - 01:18:31 Future of AI Product Management - 01:33:16 Outro - 01:35:49 💼 Check out our sponsors: Linear: Plan and build products like the best - https://linear.app/partners/aakash Miro: The innovation workspace - http://miro.pxf.io/PO4WZX Amplitude: Try their 2-minute assessment of your company’s digital maturity - https://bit.ly/4hl25RG 👀 Where to find Pawel: LinkedIn: https://www.linkedin.com/in/pawel-huryn Newsletter: https://www.productcompass.pm YouTube: https://www.youtube.com/@pawelhuryn 👨‍💻 Where to find Aakash: Twitter: https://www.twitter.com/aakashg0 LinkedIn: https://www.linkedin.com/in/aagupta/ Instagram: https://www.instagram.com/aakashg0/ Transcript: https://www.news.aakashg.com/p/complete-course-ai-product-management 🔔 Subscribe and like the video to support our content! 🔑 Key Takeaways 1. Prompting isn’t a Trick, it’s the Product. Prompting isn’t something you tack on at the end…It’s a core part of how the product works.Well-structured prompts completely change the quality of output. It’s basically the UX layer for LLMs. Your goal isn’t to outsmart the model but to teach it how to behave with clear, repeatable instructions. 2. RAG is How You Stop Hallucinations And Keep Your Product Fresh! Instead of cramming everything into the model or relying on fine-tuning, Retrieval-Augmented Generation (RAG) lets you pull in the right context when you need it. For example, he used it to pull product changelog data and get accurate responses… Without needing the model to already “know” that info. If your product updates often, RAG keeps the AI current without hardcoding anything. This is how you reduce hallucinations and keep things adaptable. 3. Most PMs fine-tune When They Should just Prompt Better. He has seen this mistake countless times: PMs reach for fine-tuning too early. He showed a side-by-side of zero-shot, few-shot, and a fine-tuned model.All summarizing a product dashboard.The few-shot prompt actually did better than the fine-tuned version. Most PMs go straight to fine-tuning, but with the right prompt structure, you can get 95% of the result!And it’s way faster, cheaper, and easier to maintain. 4. An AI Agent Is Just a Pipeline That Thinks The term “agent” gets thrown around a lot, but under the hood, it’s a system that can think: Intent classification, tool selection, execution logic, and error handling. If you don’t design for that structure, your agent becomes unpredictable fast. The real magic happens when you coordinate its behavior with reliable systems thinking and that’s your moat! 5. AI PRDs Need a New Language! Traditional PRDs were built for deterministic systems.You specify inputs, define expected outputs, and call it done. You’re not writing “requirements”, you’re writing intent, behavior, and expected failure modes. Here’s how to write PRDs for AI products: → Include structured prompts, not just user flows→ Provide real input/output examples→ Define what “acceptable variance” looks like→ Plan for fallbacks, retries, and recovery UX Most importantly: You’re not managing the model, you’re collaborating with it. And if your PRD doesn’t reflect that dynamic, your product will feel brittle, unpredictable, or worse… totally misaligned with user needs. #podcast #productmanagement #ai 🧠 About Product Growth: The world's largest podcast focused solely on product + growth, with over 167K listeners. Hosted by Aakash Gupta, who spent 16 years in PM, rising to VP of product, this 2x/ week show covers product and growth topics in depth.

Pawel HurynguestAakash Guptahost
Apr 22, 20251h 36mWatch on YouTube ↗

CHAPTERS

  1. Why AI Product Managers are in high demand (and highly paid)

    Aakash and Pawel open by discussing the rapid growth of the AI market and why AI PM compensation is often higher than traditional PM roles. The core reason: AI PMs must blend strong business/product judgment with enough technical fluency to collaborate effectively with engineers in a fast-moving space.

  2. Prompting as a foundational AI PM skill: context + structure beats casual prompts

    Pawel demonstrates a high-quality prompt aimed at uncovering hidden assumptions in continuous product discovery. He emphasizes giving rich context and explicitly specifying the thinking steps and outputs to produce more reliable, actionable results from LLMs.

  3. Prompting “hacks” and best practices PMs can apply immediately

    They expand from the demo into repeatable prompting techniques that improve output quality and reduce bias. Pawel shares practical tactics like setting success criteria, using templates, iterating, and even adding “rewards” to increase detail in responses.

  4. AI PRDs: aligning strategy and avoiding “AI for AI’s sake”

    Pawel explains what changes in an AI PRD versus a standard PRD, developed with input from an OpenAI product lead. The PRD’s main job is organizational alignment and surfacing assumptions—especially important amid AI hype and pressure to build AI features without a business case.

  5. AI PRD template walkthrough: market, users, value, defensibility, scope, GTM

    They outline the major sections of Pawel’s AI PRD template and how each section helps validate the opportunity and execution approach. The template emphasizes segmentation, value proposition clarity, sustainable competitive advantage, evaluation considerations, and go-to-market planning.

  6. Fine-tuning demo: when prompts are too heavy and style/behavior must be consistent

    Pawel shows why repeatedly injecting long “style prompts” is costly and unreliable, using a clone example. He demonstrates supervised fine-tuning on OpenAI (Yoda-style dataset) to internalize behavior into model weights, enabling cheaper models and shorter prompts for repeated use cases.

  7. Evaluating fine-tunes: reading training/validation curves and knowing when to stop

    They interpret the fine-tuning dashboard outputs, including training loss trends and validation checkpoints after each epoch. The takeaway is practical: you often reach “good enough” quickly, and extra epochs can be wasteful or even harmful.

  8. RAG explained: answering from many documents without paying to stuff them into prompts

    Pawel builds a RAG chatbot over a Google Drive folder of articles to demonstrate retrieval at scale. He explains why you can’t just paste lots of documents into the prompt (cost and scale) and introduces vector databases as the practical solution.

  9. Building a RAG pipeline live: Google Drive → n8n workflow → Pinecone embeddings

    They implement an ingestion workflow: detect new files in a Google Drive folder, download them, chunk them, embed them, and store them in Pinecone with metadata. The live build includes real-world iteration/debugging, reinforcing how these systems are developed in practice.

  10. RAG query endpoint: webhook retrieval + aggregation + LLM response with citations

    Pawel creates a webhook endpoint that takes a user query, retrieves top-k similar chunks from Pinecone, and sends them to an LLM to generate an answer grounded only in retrieved context. The response includes cited sources using stored metadata (document names).

  11. Shipping a simple RAG chatbot UI quickly with Lovable (and common pitfalls)

    They use Lovable to generate a lightweight chatbot frontend that calls the webhook and renders responses. The demo highlights how fast prototyping can be, while also showing typical issues like inconsistent markdown rendering and the need for iterative prompting/examples.

  12. MCP in practice: agents reading Figma and creating Jira epics/stories automatically

    Pawel introduces MCP (Anthropic’s standard) as a way for agents to discover and use tool capabilities via consistent interfaces. He demos Claude connecting to Figma and Jira via MCP servers to generate epics and INVEST-style user stories with acceptance criteria from a design file.

  13. AI agents and orchestration: deep market research + tool-using agents in n8n

    They close with AI agents as the top of the “AI PM pyramid,” including a deep market research agent that plans tasks, delegates to sub-agents, gathers sources, and synthesizes product-focused insights. Pawel also shows how n8n agents can use tools (email, calendar, Jira) to execute multi-step actions, emphasizing orchestration patterns and common setup gotchas.

Get more out of YouTube videos.

High quality summaries for YouTube videos. Accurate transcripts to search & find moments. Powered by ChatGPT & Claude AI.