Become an AI orchestrator: How to automate daily workflows with multi-agent systems

Apr 13, 2026 · Lenny's Podcast
🎧 PodShort 25 min squeezed to 2 AI SprinklerAS AI / ML
Episode artwork
JJ Engler
AI Power User at 10x
Lenny's Podcast
25 min squeezed to 2
Full episode from Lenny's Podcast
Quotable Moments

One of the things I love about having this project system, you're orchestrating now. Even though you might not be a developer, you're now an AI orchestrator. You can run many agents at once and have this top-level view to be able to just quickly see which agent needs your attention and just give them those permissions or to pop back in.

JJ EnglerAI Power User at 10x
00:00

If you don't tell AI what success is to you, AI doesn't know what success is for you. So I always love including good examples and bad examples.

JJ EnglerAI Power User at 10x
22:00

If you're a product manager, build it and like put your boss in there, your engineering partner, and your customer and say, every PRD, review from these three points of view and give me feedback.

JJ EnglerAI Power User at 10x
17:30
Key Insights
  • AI orchestrators can run many agents at once and have a top-level view to quickly identify which agent needs attention and grant permissions.
  • Co-work allows users to connect their business tools like Slack, Gmail, and Notion, enabling AI to process information and assist with tasks more effectively.
  • A 'project' in Co-work is essentially a folder on your computer where all files and context for a specific task are stored, allowing AI to access and work with that information.
  • By feeding AI a series of your past communications (e.g., 100 emails), it can learn and perfectly match your writing tone and style for future drafts.
  • Sub-advisory skills in Co-work allow you to launch multiple sub-agents, each with a distinct persona (e.g., AI builder, security expert, customer), to review work from different perspectives and provide honest feedback.
  • Automating tasks like a 'morning debrief' can prepare a daily plan by analyzing emails, Slack, and calendars, ensuring you're aware of priorities and prepped for meetings.
  • To get the best results from AI, it's crucial to provide clear examples of what constitutes 'good' and 'bad' output, as AI needs to understand your subjective preferences.
  • Co-work projects can be used for personal tasks like house maintenance or wedding planning, not just work-related activities, by organizing relevant information and automating processes.

RevBots.ai View:

  • AI Sprinkler teams bolt on tools like Co-work without full process redesign.
  • Custom sub-agents mimic the ARM stage's true AI orchestration but lack full integration.
  • Personal productivity hacks don't address SaaS Hoarder's tool sprawl problem.
  • Requires manual feedback loops vs ARM's autonomous learning systems.
🎧Full Episode:Lenny's Podcast →