Your AI Agents Need a Control Layer. Here's Why (with Logan Kelly)

Most founders building with AI agents are thinking about one thing: functionality. Can it do the thing? Can I automate this? Can I scale that?What they're not thinking about is what happens when it goes wrong.Logan Kelly is the CEO of Waxell AI; a governance and observability platform built to control agents running in production. He's an operator turned founder who learned firsthand that autonomy without guardrails doesn't scale. So he built the infrastructure to fix it.In this episode of Growth Department, Logan breaks down why governance is the missing layer in most agentic AI stacks; and why putting it in place before your first incident is the only move that makes sense.You'll learn:The difference between observability and governance; and why you need bothHow AI agents can be hacked through context poisoning and rug pull attacksWhy your agentic chatbot is an attack surface most founders aren't protectingHow to prevent a thousand dollar API bill from appearing overnightWhy governance should sit above your agents, not inside themHow to scale agent infrastructure without replacing your peopleWhether you're a founder just getting started with AI agents or a leader ready to operationalize them across your business, this episode will change how you think about what it means to deploy AI safely.🔗 Follow Logan and learn more about Waxell → waxell.ai LinkedIn: https://www.linkedin.com/in/logankkelly/Check out his favorite podcast: Behind the Bastards🎧 Subscribe to Growth Department on Apple Podcasts, Spotify, and YouTube. 🌐 More from Growth Department → growthdepartment.comGrowth Department slows the conversation down so founders and business leaders can go further, faster.If this episode helped you think differently, follow the show and share it with someone building alongside you.