Interested in being a guest? Email us at [email protected]

AI is moving faster than any wave we’ve seen—and the hidden cost is complexity. We sit down with New Relic’s leadership to unpack how observability must evolve to keep up with agents, vibe coding, and probabilistic systems that don’t fail like traditional software. From the early SaaS days to hyperscale cloud and now AI-native architectures, we connect the dots on where value breaks, why pilots get stuck, and what changes when real users meet real AI in production.
We challenge the headline that “95% of GenAI projects fail” and explain the real blockers: adoption, verification, and integration into messy enterprise workflows. Then we dive into the rise of AI agents—microagents and nanoagents coordinating around-the-clock—to offload repetitive tasks and accelerate delivery. With that power comes a new ops reality: tracing prompts and tool calls, monitoring hallucinations, enforcing safety policies, and tying every hop back to user and business outcomes. You’ll hear practical guardrails for vibe coding, including evaluation suites, prompt versioning, canary releases, role-based access, and real-time model quality monitoring.
Finally, we look at outages through a clear lens: even the smartest teams can’t outpace system complexity without better visibility and faster remediation. New Relic’s two-pronged roadmap—observability for AI and AI for observability—aims to give teams both the data and the decisions: model-aware telemetry, agent state insights, automated incident summaries, and recommended fixes that protect revenue and trust. If you’re planning for 2026, take this as your cue to embrace the era with discipline: ship agents, instrument everything, and let observability become the backbone of safe, scalable AI.
If this conversation helps you think sharper about AI in production, follow the show, share it with your team, and leave a quick review so others can find it.

Support the show

More at https://linktr.ee/EvanKirstel