How experienced product managers use ChatGPT
0 likes17 views
The document outlines a presentation on using AI, specifically ChatGPT, to enhance productivity in product management, led by Tal Raviv. It shares insights on utilizing language models effectively, emphasizing the importance of context and the advantages of speech-to-text technology. The session includes live demos and strategies to improve product artifacts and personal productivity systems.
How experienced product managers use ChatGPT
- 1. 🎞 Live prompt demos recording is here 🤖 Each prompt from recording is in the speaker notes below
- 2. Practical skills and tangible tools to accelerate your career immediately How experienced product managers use ChatGPT Tal Raviv AI PM, Early @ Riverside, Patreon, Appsflyer, Wix Welcome
- 3. Let’s gooooo 2 min: Icebreaker 4 min: I was a hypocrite → 2 LLM insights 14 min: Live demos 1 min: My course Open Q&A (will stay until my hard stop in ~50 min)
- 4. 50 seconds: Getting to know you We’ll use slido
- 5. 70 seconds: Tell me—why are you here? Share in the chat: what’s your biggest bottleneck you hope LLMs can help with? Optional: What have you tried so far and how did it go?
- 6. On one hand, I lead multiple AI products used by tens of thousands of creators, celebrities, huge media co’s (Marvel, Google, Spotify, Netflix, CNN, NBC, NYTimes) On the other hand, I didn’t use ChatGPT almost at all in my day to day. I worried it would: ● Make me sound average ● Dull my storytelling & voice ● Take more time to fix and edit than to write it Another fear that I had was that since writing = thinking, by not doing the writing myself, I would have fewer insights. All these assumptions proved wrong when I faced… As an AI PM—I didn’t use AI at work 6
- 7. 😱My team kicked off a huge initiative affecting every corner of the product. 😱😱 My eng lead wanted to move to a strict Scrum model* and wanted very elaborate user stories. (Links to Figma weren’t enough anymore) 😱😱😱 Figma AI plugins were creating more work than they were worth 😱😱😱😱 I needed to create dozens of jiras each with: acceptance criteria, user stories, objectives, etc. that would take me forever My breaking point 7 *Many thoughts on scrum, saving for another time Step 1: Whine & complain to Oleksii (my eng team lead) Step 2: Oleksii shows me how to actually use ChatGPT Step 3: I use it forever for all my product artifacts (go to market, PRD, email sequences, slack blasts)
- 8. Insight 1: Shower LLMs with context 8
- 9. 1. LLMs excel at synthesizing, inferring, formatting, & articulating 2. LLMs blind spots are unwritten things the world hasn’t yet seen 3. Great PMs are massive sources of context on customers, market, strategy, design principles, focus guardrails, narrative… LLMs are average, but you’re not 9
- 10. Insight 2: Use speech-to-text!!!!!!!111 10
- 11. 1. Incredibly accurate speech-to-text with Whisper (BetterDictation, SuperWhisper, MacWhisper, ChatGPT desktop) 2. Dictating is awkward. Whoever gets over it will have a massive AI advantage. 3. Verbalizing leads to new insights (uses different cognitive processes—we’ve all experienced this) PMs can talk 11
- 12. Ever since Whisper came out, I use it for: 1. Slack messages 2. Email replies 3. Writing documents 4. Quickly jotting down ideas 5. Debriefing/summaries after user research 6. This slide deck PS: Speech-to-text is a productivity revolution 12
- 13. Live demos We’ll use the Lime scooter app’s “park without GPS” as our example feature: 13
- 14. Live demo: Opportunity doc (4o) 14
- 15. Live demo: User stories (4o) 15
- 16. Live demo: Casual Jira (4o) 16
- 17. Live demo: Slack msg in my voice (o1) 17
- 18. Live demo: Blog post in my voice (o1) 18 Note: In case we don’t get to cover the longer blog post example, here's a 3-min loom of me writing a full blog post by purely dictating into GPT-o1
- 19. Pro tips 19
- 20. 1. Choice of template and structure (e.g. the headings of your document) 2. Breakdown into chunks (user stories, sections, initiatives, phases, email sequence) 3. Unique context built around company strategy + spending time with customers PS: 1 & 2 can be done with the help of LLMs, just do it as its own step in advance, so you can inspect Set up LLMs for success by having this ready: 20
- 21. 1. Raw support tickets 2. Raw user research & sales conversations 3. Raw qualitative survey feedback 4. Competitive research, dogfood it yourself 5. Iterate/tinker with your own data charts and segments Stay competitive—don’t use AI here: 21 6. Watch the gong call excerpt, hearing the customer's voice 7. Parse the actual Reddit thread about your product 8. View the Fullstory playlist 9. Review slack feedback from Sales/CSM (don't stop at slack AI channel summaries) 🖋 Use AI for producing, not consuming.
- 22. Save $138 with code PETERYANG829 The QR code above automatically applies the $138 discount Build Your Personal PM Productivity System Oct 28-Nov 10 • 🏆 Lenny’s “Top Course in Product” • Students from Microsoft, Amazon, Meta, Tiktok • 4x live sessions + 2x 90-min AMAs Email b2b@maven.com for a bulk purchase. View course syllabus: 22
- 23. Save $138 with code PETERYANG829 The QR code above automatically applies the $138 discount What I’ll send out when the recording is ready 📺 Link to Slides (prompts in speaker notes) 🎞 Video recording 📑 Links to Lenny’s top doc templates 📱 My contact info for questions Email b2b@maven.com for a bulk purchase. View course syllabus: 23
- 24. Open Q&A in chat 24