The AI Productivity Podcast Our Claude AI Employees Run The Business
24 snips
Mar 4, 2026 They describe building a 28-member AI team with persona-based agents that run hiring, thumbnails, analytics and more. Systems thinking and SOPs are framed as the secret to reliable AI work. Conversations cover persistent memory, orchestrator agents that index and delegate, and how to scale by cloning agents for parallel tasks.
AI Snips
Chapters
Transcript
Episode notes
How Larry The Orchestrator Runs Content Pipelines
- Tom Solid built a 28-member AI team starting with an orchestrator agent named Larry to delegate tasks like thumbnails and blog posts.
- Larry routes work to specialists (Pixel, Marty, Penn) and creates ClickUp tasks, mirroring a real team workflow.
Agent HR Hires New AI Specialists
- Tom created a dedicated HR agent (Nolan) so the system can hire new specialist agents on demand when gaps appear.
- Agents run a hiring and onboarding process: role profile, charter, SOPs, then integrate into the index file.
Design Agents Narrowly And Use One Orchestrator
- Keep agents narrowly focused and use one orchestrator to index and delegate so sessions stay fast and reliable.
- Split knowledge into small profile/SOP files to avoid context-window ceilings and degraded outputs.