Lenny Rachitsky · 4 min read · Gen AI Tools
🎙️ How I AI: How this PM uses Claude to handle 70% to 80% of his workday
Read the original on lennysnewsletter.com →
Daniel Blum built a self-improving AI assistant with Claude that handles 70-80% of his workday by connecting Claude to Notion, Slack, email, and other tools. He describes how architecture, ongoing context, and learning from edits let the system scale, onboarding others in 15 minutes and eventually letting him accomplish a week's work in a day.
Key Takeaways
- Architecture matters more than the AI tool itself: autonomously updating core files and connecting to the tools you already use yields scalable improvement.
- Context is ongoing: continuously feed Claude contextual data (voice memos, links, decks) and refresh it regularly to keep the AI aligned with real work.
- The system proactively identifies what it doesn't know and asks targeted questions to fill gaps.
- Self-improvement loops learn from the difference between drafts and final outcomes, enabling the AI to improve over time without requiring explicit feedback.
- Feedback telemetry and weekly improvement loops transform personal AI workflows into a repeatable, improving product for the organization.
Excerpt
How I turned Claude into a self-improving PM assistant | Daniel Blum (PM, Melio) Listen now on YouTube • Spotify • Apple Podcasts Brought to you by: Optimizely —Your AI agent orchestration platform for marketing and digital teams Jira AI SDLC —Get your tokens’ worth with Jira Daniel Blum is a product manager at Melio who has built a self-improving AI system that now handles 70% to 80% of his workday. In this episode, he breaks down how Claude and Cowork manage his Notion board, prepare him for the week, scan Slack and email for important context, and learn from…