The Vibe PM – Détails, épisodes et analyse

The promises of AI in software are abundant, and for something that’s supposed to save you time, it’s overwhelming to learn which tools will give you the most out of AI.

The Vibe PM podcast, hosted by Dave Killeen, features use cases and demos of the latest AI tools and techniques. Dave goes in deep, so you don’t have to.

Join Dave to get the AI insights - curated and tested - to keep you ahead on AI and on top of what matters most in your day.

Quick tips. Smart moves. Big vibes.

Presented by Pendo

Learn more at pendo.io/
Follow Dave Killeen on LinkedIn

You're typing 50 words per minute, but your brain can move at 200+. Ready to ditch the keyboard bottleneck and gain up to nine hours back in your week?

In this debut episode of The Vibe PM, Pendo Field CPO EMEA and AI enthusiast, Dave Killeen, dives into the ultimate productivity hack: turning your voice into action with cutting-edge AI speech-to-text tools. 

From raw dictation speed to intelligent voice modes, Dave demos five increasing levels of speech-to-text AI integration that help product managers, tech leaders, and any busy professional get more done in less time.

Chapters:

00:00 Intro

00:20 Why voice beats typing (the math)

01:30 Level 1: WhisperFlow — raw speed dictation

02:40 Level 2: Superwhisper — custom voice input modes

04:45 Level 3: Walking chats with LLMs (mobile)

07:10 Level 4: “Magic Word” batching → exec doc

09:30 Level 5: Limitless.ai pendant — ambient AI

11:10 Wrap

Tools Mentioned:

• WhisperFlow (speech‑to‑text speed)

• Superwhisper (voice modes for Docs/Gmail/Slack)

• ChatGPT/LLMs on mobile (voice conversations)

• Limitless.ai pendant (hands‑free capture)

  
  

No fluff, no hype - just Dave sharing the practical ways to work faster to keep you vibing throughout your day.

The Vibe PM: Quick tips. Smart moves. Big vibes.

Presented by Pendo

Connect with Dave Killeen on LinkedIn | Explore more at pendo.io

You know the feeling - you’ve been asked to create a presentation for a conference or a customer meeting, and you find yourself frozen, staring at a blank slide. The ideas are in your head, but getting them into a story and a deck that resonates is frustrating and time-consuming.

In episode two of The Vibe PM, Dave Killeen, Field CPO EMEA at Pendo reveals his two-part AI workflow for creating memorable presentations, including the one he gave at ProductCon London in early 2025. 

First, Dave shows how to use the power of voice to turn brain dumps into structured, story-driven speaker notes. 

Second, Dave demonstrates how to transform those speaker notes into on-brand, professional slides using Gamma, one of Dave’s favorite AI presentation tools.

And when the presentation is done, Dave shares a bonus tip for using these tools to really connect with the audience.

At the heart of it, AI isn’t here to replace your creativity - rather, it amplifies it. These tools don't write the story for you; they handle the time-consuming parts so you’re free to do the real work - thinking deeply about your message, your audience, and the impact you want to make.

Chapters:

00:00 Intro: Why decks kill PM productivity

01:16 Step 1: The "Magic Word" brain dump (voice → Claude)

03:35 Step 2: Gamma.app turns notes into branded slides

05:35 Advanced prompt hack for diverse, bias-free images

07:02 The Golden Asset (spoiler: it's not the deck)

08:13 Wrap: From one conversation to infinite assets

Tools Mentioned:

• Claude (mobile voice mode for structured brain dumps)

• Gamma.app (AI presentation builder with brand controls)

• Advanced image prompting techniques

The Vibe PM: Quick tips. Smart moves. Big vibes.

Presented by Pendo

Connect with Dave Killeen on LinkedIn | Explore more at pendo.io

Here's the uncomfortable truth: if you're getting mediocre output from AI, it may not be the technology; it may be you. 

In this episode of The Vibe PM, Dave Killeen demos how to switch up that dynamic using the Claude Console's Prompt Improver - a tool that transforms your vanilla AI prompts into detailed, results-oriented prompts that actually get you strategic, usable results faster. 

This takes practice, but once you begin focusing on the prompt, you’ll achieve dramatically better results with your work, helping you transition from a good PM to a great one. 

Chapters:

• 00:00 Why your AI prompts are failing you 
• 02:07 Claude's Prompt Improver (the game-changer) 
• 02:40 Demo pt. 1 (creating a test strategy doc) 
• 03:36 Demo pt. 2 (the "magical" prompt generation) 
• 04:47 Side-by-side (basic vs. expert prompt comparison) 
• 06:05 The secret sauce (treat AI like an expert consultant) 
• 06:50 Three ways to use this in your workflow now 
• 07:46 How much does this cost? (spoiler: almost nothing) 
• 08:14 Your challenge (the one-way door decision) 
• 08:44 Wrap (get more time back in your day)

Tools mentioned:

• Claude Console (Prompt Improver tool)
• Free account access at console.anthropic.com

The Vibe PM: Quick Tips. AI Flows. Big Vibes.

Presented by Pendo

Connect with Dave Killeen on LinkedIn | Explore more at pendo.io

Are you struggling to bring your product idea to life? What about the speed in which it takes you to complete your discovery and validation cycles? Yeah, the frustration is real.

Here's the thing: If you’re using an LLM to flesh out your ideas, you've been doing it backwards.

In this episode, Pendo Field CPO, EMEA, Dave Killeen, introduces context pulling - a method that flips traditional AI prompting on its head. Instead of struggling to get anything useful out of the AI with the context you're providing, you engineer the AI to actively pull the context it needs from you. It then turns your idea into a spec'd, flow-mapped, demo-ready prototype faster than you thought possible.

This isn't theory. Dave walks you through the exact prompt, the magic commands that generate instant assets, and why answering AI questions with voice mode is the secret weapon most PMs are missing out on.

Chapters:

• 00:00 The context-pushing trap (and why your outputs aren’t useful)
• 01:20 Context pulling: flip the script, let AI lead
• 02:40 Build your Virtual PM (the prompt breakdown)
• 04:15 Magic commands that do the heavy lifting
• 05:30 Voice mode = cheat code for speed
• 06:20 Real talk: idea to prototype in under an hour

Tools mentioned:

• Claude
• Loveable
• Mobile voice mode
• Dave's exact prompt (in description)

Dave’s Prompt for Context Pulling

You are an AI agent with extensive experience in product building, product management, and product design. You excel at taking requirements and generating clear scopes, creating user flows and detailed key screen specifications.

When presented with a product idea or feature, your task is to act as a product manager and designer to help refine and develop the concept into a detailed requirements document for AI-powered coding platforms.

Proceed through these steps:

CLARIFYING QUESTIONS

After receiving a product idea, generate 8-10 important questions as a product manager to understand the requirements better. Present these in a numbered list.

SPEC DOC

When user says "spec doc", analyze previous responses to generate a product spec for so that we're in a position to design the flows after this, including:

• Product summary
• Problems solved
• Target audience
• Key features and functionality
• Success metrics

DESIGN FLOWS

When user says "design flows", create detailed user flows and screen specifications. For each screen:

• Describe layout and UI elements in detail
• List all possible user actions and outcomes
• Specify navigation flows and transitions
• Detail any validation rules or state changes

PRD

When the users says “create PRD”, create a detailed PRD that would be in enough detail for a squad to be able to pick up and run with.

After each of the above steps, include:

• Next step command prompt
• Option to revisit previous steps
• Keep responses thorough but structured and clear. Focus on requirements (what) rather than implementation (how).

The Vibe PM: Quick Tips. AI Flows. Big Vibes.

Presented by Pendo

Connect with Dave Killeen on LinkedIn | Explore more at https://www.pendo.io/https://www.pendo.io/

What if you could advance your career by building an entire application—complete with multiple modes, persistent memory, and personalized intelligence—using nothing but your voice in natural language? No code. No developers. Just words.

Most teams are still treating AI like a search bar: ask a question, get an answer, start over. They're missing the bigger opportunity entirely.

In this episode of The Vibe PM, Pendo Field CPO EMEA Dave Killeen demonstrates how to shift from transactional prompting to building prompt-based applications—sophisticated tools that remember context, ask intelligent questions, and compound value over time.

Dave's example: building a multi-mode AI Career Coach live on the show. This isn't a chatbot, it's a prompt-based application architecture designed in minutes that handles weekly check-in summaries, monthly reflections, promotion readiness assessments, and executive report generation.

The methodology is refreshingly simple: prompt the AI to serve as your personal career coach, asking you the right work value questions, load job-related context once, and consistently return to the same work value conversation throughout the year.

Chapters:

• 00:00 Introduction: Applications built with words, not code
• 02:00 Foundation: Setting up Claude Projects for persistent context
• 02:50 Architecture: Generating the multi-mode mega prompt
• 05:05 Input method: Weekly brain dumps via voice
• 06:00 The flip: When AI starts asking the questions
• 07:05 Live demo: Unstructured thoughts to polished weekly report
• 08:20 The four modes: Beyond basic check-ins
• 09:00 Your blueprint: Three steps to build prompt-based apps
• 10:00 The compounding advantage: Why consistency wins

Tools mentioned:

• Claude Console (prompt generation)
• Claude Projects (persistent context storage)
• Mobile voice mode

The Vibe PM: Quick tips. AI flows. Big Vibes.

Presented by Pendo

Connect with Dave Killeen on LinkedIn | Explore more at pendo.io

Most of us have scattered notes, ideas, and reference articles - some on sticky notes, others in Google docs, but this isn’t an adequate system. That’s where a Personal Knowledge Management (PKM) system is valuable. A PKM system collects, organizes, and retrieves information, turning it into usable, compounded knowledge.

But knowledge without action will not lead to future success. You need a Personal Operating System (POS) for managing how you work - your daily routines, goals, priorities, tasks, and career growth. 

The two systems work together. Your PKM stores what you know, and your POS uses that knowledge to run your day, your week, your quarter, effectively.

Until recently, building both systems required massive manual discipline. AI changes that. With AI tools, you can now turn scattered information into a coherent PKM and use it to build a POS that gets smarter every single day.

In this episode, Pendo Field CPO EMEA Dave Killeen shows you how to build both systems, no coding required. Using Claude and Cursor, he built both systems that learns through compound engineering, where every interaction makes the system exponentially smarter.

On Day One, the AI output is generic, but by Day 30, it has learned how you operate, what you care about, and how to keep you focused.

Dave's giving you the Starter Kit below. Just clone the repo, follow the prompts, and the system begins building its structure for your role in minutes. 

Chapters:

• 00:00 Intro: Why Personal Knowledge Management (PKM) systems become unwieldy over time
• 00:47 The Claude Hooks unlock: Systems that learn automatically
• 02:10 Compound engineering explained: Every feature makes the next easier
• 03:16 Introducing your Personal Operating System (POS)
• 04:17 Demo begins: Setting up the Dex repository
• 05:14 Inside Cursor: How the system components work together
• 06:15 Running the setup command: Five-minute onboarding walkthrough
• 07:17 Choosing your role: Selecting from 30+ professional templates
• 09:30 Manual versus automatic meeting processing options
• 10:04 Watch the system create your personalized folder structure
• 11:15 Next steps: Clone, onboard, and create your first projects
• 13:22 The competitive edge: Transformational impact, not incremental

Links to Dave’s Starter Kit

• Dave’s article on how to build a personal operating system
• Github

Tools mentioned

• Claude (with Claude Hooks for automation)
• Cursor IDE
• Granola (meeting transcription with MCP integration)
• MCP (Model Context Protocol) for connecting tools

The Vibe PM: Quick tips. AI flows. Big vibes.

Presented by Pendo

You've built your app's foundation in Lovable or AI Studio, and suddenly, you have multiple feature ideas flooding your brain. You likely jump into a sequential chat-based coding app to start building all these ideas, and you realize it feels painfully slow. You request a feature. You wait. It builds. You request another. You wait. It builds. Repeat. It's just become too slow.

In this episode of The Vibe PM, Pendo Field CPO EMEA Dave Killeen shows you how to move beyond sequential building to rapid iteration using parallel agents.

Dave demos the exact workflow for working with parallel agents to vibe code faster. Dave suggests starting with Lovable or Google AI Studio to get your app foundation set. Then you can explore ideas using tools such as GitHub and Cursor, and when you’re ready, fire up multiple agents simultaneously, each working on different features at the same time. The pace of progress becomes fundamentally different, which Dave calls "lightning pace" vibe coding.

So what did Dave build with his parallel agents? He built a KPI Driver Tree app in two hours on a plane. He was inspired after reading an Atlassian stat that only 12% of PMs feel connected to business outcomes. The app maps company metrics to north star outcomes, helps teams understand how their work connects to board-level dollars , and includes a feature that critiques its own strategy.

This episode isn't about the KPI tree itself - it's about showing you what becomes possible when you stop waiting for one AI agent to finish before starting the next idea.

Chapters:

• 00:00 Intro: Why most PMs can't connect work to dollars
• 00:21 Beyond Lovable: The next level of vibe coding
• 01:24 The problem with sequential AI coding
• 02:00 Why you need parallel agents to build faster
• 02:35 Demo: Lovable and Google AI Studio for foundations
• 03:30 Pushing to GitHub and pulling into Cursor
• 03:50 Inside Cursor IDE: Setting up your workflow
• 04:15 GitHub trending: Forking existing projects to start fast
• 05:11 Agent mode in Cursor: Building without seeing code
• 05:45 Firing up multiple agents simultaneously
• 06:30 The fastest way forward: Lightning-pace vibe coding
• 06:34 The KPI driver tree demo begins
• 07:00 The Atlassian 12% stat: Why strategic alignment matters
• 07:55 Live demo: Generating a KPI strategy for Trade Me
• 08:30 Understanding the three vectors: Revenue, cost, risk
• 09:00 Red team analysis: AI critiques AI
• 09:45 "Your KPI structure has structural flaws"
• 10:30 "You're incentivizing a landfill" (AI roasts the strategy)
• 11:30 Deep dives on metrics and benchmarks
• 12:00 Building a sparring partner for your work
• 12:44 Wrap: Moving to the next level of vibe coding

Tools mentioned:

• Lovable (foundation building)
• Google AI Studio (app scaffolding)
• GitHub (version control)
• Cursor IDE (parallel agent workflows)
• Alternative IDEs: Windsurf, Gravity

The Vibe PM: Quick tips. AI flows. Big vibes.

Presented by Pendo

Connect with Dave Killeen on LinkedIn | Explore more at pendo.io

You know the feeling. You open that resume you haven't touched in months or years, stare at it, and think: "How do I describe what I’ve accomplished? How do I describe that project? How do I quantify the impact I’ve made?"

Here's the brutal truth: we're terrible at remembering our own achievements. We actually do amazing work, but when it's time to write it down, our minds go blank. Then we ask ourselves: How can AI help make my resume better? How can AI help my LinkedIn profile better?

The big idea today, though, isn't actually about the resume. In this episode, Pendo Field CPO EMEA Dave Killeen shows you the power of “burger prompting” in the LLM. It’s time to move from the old way of LLM prompting, where you're guessing what context the AI needs. Instead, you’ll learn how to build and improve prompts so the AI extracts the information it needs from you to create a usable output.

Dave demonstrates this through six-phase structured prompt that instructs the AI to conduct a structured interview to surface your brilliance and specific achievements so you can have an outstanding resume.

Chapters:

• 00:00 The resume problem (why we all go blank)
• 01:00 Flipping the script (AI interviews you, not the other way around)
• 02:00 The six-phase structure (from upload to LinkedIn profile)
• 02:30 Phase 3: Achievement extraction (where the magic happens)
• 03:30 Live demo: Alex's mediocre resume gets transformed
• 05:00 How the prompt refuses vague answers
• 07:00 The full prompt structure breakdown
• 08:00 Building burger prompts with Claude Console
• 09:00 Wrap (the bigger idea: context pulling in action)

Tools mentioned:

• Claude (structured interview process)
• Claude Console (Prompt Improver for building these workflows)
• Dave's exact resume builder prompt

You'll walk away with:

• A six-phase prompt that interviews you to extract forgotten achievements
• The methodology that refuses vague statements and demands quantifiable results
• A system that generates both your resume and LinkedIn profile from one conversation
• Understanding of how to build "burger prompts" that pull context from you

The Vibe PM: Quick tips. AI flows. Big Vibes.

Presented by Pendo

Connect with Dave Killeen on LinkedIn | Explore more at Pendo.io

Model Context Protocol (MCP) is becoming one of the most important tools for getting more value from AI workflows. In this episode, Pendo Field CPO, EMEA Dave Killeen explains what MCP is, why it matters, and how it helps organizations turn their scattered, siloed data into synthesized, useful insights that no single tool could offer on its own.

MCPs connect your AI assistant to multiple tools and data sources, so instead of jumping between dashboards, you can ask questions in natural language and get answers you really need.

There are several MCPs available, and in this episode, Dave Killeen walks through four real-world scenarios using Pendo’s MCP to demonstrate the value MCPs have on a PM's day-to-day work.

Chapters:

• 00:00 Intro: Why MCP is everywhere right now
• 00:30 What makes MCP different from traditional integrations
• 01:10 The Pendo MCP: What it connects to
• 01:41 Demo 1: Account review prep (QBR in 30 minutes)
• 02:55 Demo 2: Portfolio review (which accounts need attention)
• 04:05 Demo 3: Feedback themes across entire customer base
• 04:45 Demo 4: From insights to executive slides
• 06:22 Wrap: What MCP changes about how you work
• 07:05 Where to find MCP servers (smithery.ai)

Tools mentioned:

• MCP (Model Context Protocol)
• Pendo MCP (product engagement, feedback, user behavior data)
• Claude (AI with MCP integration)
• Smithery.ai (MCP server directory)
• Anthropic's built-in slides generator

The Vibe PM: Quick tips. AI flows. Big vibes.

Presented by Pendo

Connect with Dave Killeen on LinkedIn | Explore more at pendo.io