AI agent management is the new core competency for revenue teams

Aug 24, 2026 · Lenny's Podcast
🎧 PodShort 23 min squeezed to 2 Ai sprinklerAI AI / ML New
Episode artwork
Ryan Carson
Solo Founder at Untangle
Lenny's Podcast
23 min squeezed to 2
Full episode from Lenny's Podcast
Quotable Moments

All of us have to uplevel our ability to manage agents. Like that is our job. How do you effectively manage 10, 20, 100, 500 employees?

Ryan CarsonSolo Founder at Untangle
03:00

I really don't ship more than I think the market wants. Like I really try to constrain my output, not on quality, not on bugs, but like because I don't I don't think I get multiples of quality off of multiples of output.

Ryan CarsonSolo Founder at Untangle
10:10

I think people are not getting out of their chair enough and actually talking to real people. So people, go talk to humans for real and build stuff they need.

Ryan CarsonSolo Founder at Untangle
11:00
Key Insights
  • The future of engineering work is largely with cloud agents, moving away from local machine-based development.
  • The primary job for professionals in the AI age is to effectively manage a multitude of AI agents, akin to managing a large team of employees.
  • AI agents can be effectively used for operational tasks beyond just coding, such as managing deal desks, customer quoting, and documentation.
  • Decision fatigue can be used as a prioritization heuristic; if a feature is easily abandoned due to fatigue, it might not be truly important.
  • Getting out of the chair and talking to real customers is crucial for finding product-market fit, as AI alone cannot create markets for products.
  • Using AI agents in public Slack channels can be an effective change agent move, making real-time coding work transparent and shifting perceptions of AI capabilities.
  • Hiring processes can be revolutionized by having candidates record videos of themselves managing AI agents to build features, providing a more effective assessment than traditional methods.
  • While Claude is excellent for creating design systems and decomposing design elements, it struggles with actual design implementation compared to tools like Codex.
Metrics Mentioned
  • $5,000 to $20,000 per month (Ryan's spending on Devin for cloud agent usage before securing credits.)
  • 40 PRs a day (Ryan's current rate of shipping Pull Requests.)
  • 2,000 unread emails (The number of unread emails Codex processed for Ryan, highlighting its efficiency in email management.)

RevBots.ai View:

  • AI Sprinkler teams must adopt agent management as a core discipline, not just tool usage.
  • Tab Hopper founders can leapfrog by using AI for ops but risk missing human feedback loops.
  • ARM-stage orgs will treat AI agents as team members with clear roles and workflows.
  • SaaS Hoarders face cost spikes without agent orchestration (see $20k/month Devin bills).
🎧Full Episode:Lenny's Podcast →