Are Your Apps Ready for AI Agents?

In this episode of The Tech Trek, Amir sits down with Reed McGinley-Stempel, co-founder and CEO of Stytch, to explore what it means for applications to be agent ready. With the rise of agentic AI—intelligent systems that can take actions on behalf of users—the landscape for SaaS and consumer-facing apps is rapidly evolving. Reed breaks down the core concepts around agent integration, including how apps must prepare to serve not just human users but also AI agents acting on their behalf. They discuss the key challenges companies face: earning user trust, managing consent and privacy, and building in human oversight to minimize costly mistakes. Using real-world examples like coding agents and calendar tools, Reed illustrates how agent adoption succeeds where there's low friction and built-in validation. He also dives into the double standard AI faces, and why even psychologically, humans might need a "human in the loop" long after AI is capable of operating on its own. If you're building applications or thinking about AI integrations, this is a forward-looking conversation you won't want to miss. 🧠 Key Takeaways What “Agent Ready” Really Means: Apps must now prepare for a world where both humans and AI agents interact with them—sometimes autonomously. Balancing Trust and Control: Consent, data privacy, and human-in-the-loop confirmations are key to gaining user trust in AI agents. Coding Agents as the First Wave: Software development is a prime use case for agent adoption, thanks to built-in validation workflows and low user friction. Why Mistakes Hit Harder with AI: Users hold AI to a higher standard than humans—especially when the cost of fixing AI mistakes causes more mental fatigue than doing it manually. The Psychological Role of Humans: Even as agents improve, a “human in the loop” may remain necessary just to reassure users, much like early elevator operators. ⏱ Timestamped Highlights 00:34 – What Stytch does: An API-first identity platform for customer apps. 01:24 – What it means to be “agent ready” in 2025. 04:29 – The 2 major user concerns: data privacy and efficacy. 08:16 – The risk of losing touch with the end user in agent-driven workflows. 11:03 – Why coding agents gained early traction: low friction + strong validation. 15:58 – Users expect more from AI than junior engineers—sometimes unfairly. 20:23 – How agent workflows challenge traditional notions of data consent. 24:17 – The future of human-in-the-loop: functional now, psychological later. 💬 Notable Quote “Humans hate friction and they hate mistakes. Agents help reduce friction—but only if they don’t make the kind of mistake that breaks trust.” – Reed McGinley-Stempel 🔗 Resources Mentioned Stytch: Identity infrastructure for modern apps 🚀 Career Tips (From the Episode) If you're an engineer, expect your role to shift toward problem solving, not boilerplate coding. When working with agents, focus on building validation steps into your workflows—they're key to adoption and trust. Product managers and designers should prioritize consent UX and asynchronous confirmations to balance automation with user control.