AI used to wait for instructions. Now, it doesn’t always ask. In this opening episode, we explore the rise of agentic AI systems that don’t just respond to input, but take initiative, set goals, and act on their own. We break down what agentic really means, why it’s different from traditional automation, and what kinds of design and trust challenges this shift introduces. Along the way, we look at how this plays out in tools that summarize, schedule, and trigger real-world workflows — and why autonomy sounds good in theory but gets messy fast. Whether you’re building LLM-powered copilots, evaluating autonomous workflows, or just trying to keep your incident response human-aware, this is the groundwork you’ll need for what’s coming next.

Sid Verma is a Developer Advocate at PagerDuty, where he helps developers optimize their workflows and implement scalable solutions. With a deep background in observability, DevOps, and enterprise open source technologies, Sid is passionate about empowering teams to innovate and improve their incident management processes. He’s also a tech enthusiast with experience in vector databases, security APIs, and more.