AI agent orchestration unlocks niche business automation

Jul 6, 2026 · Lenny's Podcast
🎧 PodShort 35 min squeezed to 3 AI SprinklerAS AI / ML New
Episode artwork
Alessio Finelli
Founder & Co-host at Kernel Labs & Latent Space Podcast
Lenny's Podcast
35 min squeezed to 3
Full episode from Lenny's Podcast
Quotable Moments

All of that work now can easily be automated, you know, even with just with the Magic Glasses or something else. And so you're kind of helping actually even at a small scale you can get a lot of leverage out of it.

Alessio FinelliFounder & Co-host at Kernel Labs & Latent Space Podcast
02:46

The whole process is super inefficient because people are like searching each card manually... You can actually use AI to save clock time for real people by doing these things autonomously.

Alessio FinelliFounder & Co-host at Kernel Labs & Latent Space Podcast
00:54

I think the reality is like, how do you build tools for it to be more effective? And that's kind of like one of the main things also Kernel Labs has been working on.

Alessio FinelliFounder & Co-host at Kernel Labs & Latent Space Podcast
01:20:06
Key Insights
  • AI facilitates significant quality of life improvements for small businesses by automating historically inefficient processes, such as manual inventory management for fish delivery services.
  • AI can automate complex tasks for trading card businesses, including tracking PSA certificates, pricing cards in real-time on platforms like eBay, and managing inventory valuation for optimal buying decisions.
  • The speaker found greater success by shifting from simply prompting an AI agent to becoming an 'agent manager' that orchestrates multiple agents and manages their workflows, often leveraging cloud infrastructure for persistence and scale.
  • True effectiveness with AI, especially for longer-running and complex tasks, comes from building robust tooling and frameworks around LLMs, rather than solely relying on the raw power of the models themselves.
  • Explicit specifications and structured workflow definitions (like Open AI's Symphony's markdown files) are crucial for guiding AI agent behavior and ensuring predictable outcomes, preventing agents from introducing 'weird' or unintended actions.
  • AI excels at handling highly heterogeneous and unstructured data that was previously impossible to scale with traditional software, unlocking new business opportunities in niches like trading cards, vintage clothing, and other 'physical world' integrations.
  • AI can act as a 'context offloader' or 'safety net' by automating mundane checks, such as constantly monitoring emails for important responses or managing personal financial accounts, thereby reducing human stress and freeing up cognitive capacity.
  • Orchestration platforms like Symphony provide a complete historical record of each task, including initial specifications, work-in-progress notes, and rework iterations, which is invaluable for debugging, learning, and continuously improving agent directives.
Metrics Mentioned
  • Over a million developers (The number of developers using Firecrawl, including the hosts.)
  • 10,000 free credits (A promotional offer for Firecrawl's web data API.)
  • 221 million tokens (The token cost for a single, complex AI engineering task (deploying a self-hostable project).)
  • 600 books (The approximate number of books cataloged in the host's house using AI, translating to over 100 books per person.)

RevBots.ai View:

  • AI Sprinkler stage: bolting AI onto legacy processes creates pockets of efficiency.
  • ARM potential: agent orchestration hints at true workflow transformation.
  • Tab Hoppers could use AI for manual task relief but lack integration.
  • SaaS Hoarders risk adding more disjointed AI tools without frameworks.
🎧Full Episode:Lenny's Podcast →