ABOUT THIS EPISODE
AI coding tools are generating more code than ever, but enterprise engineering teams are not necessarily delivering software faster.
In this episode of Drag & Drop, Feroz Khan sits down with Martin Reynolds, Field CTO at Harness, to explore the AI velocity paradox and what it means for software delivery, DevOps, AgentOps, platform engineering, and AI agents in production.
Martin explains why building an AI agent can be relatively easy, but deploying one safely requires a completely different level of governance, identity, testing, observability and accountability.
The conversation explores why code is no longer the main bottleneck, why traditional delivery processes are struggling to keep pace with AI-generated output and what engineering organizations need to operate increasingly autonomous systems safely at scale.
In this episode, you’ll learn:
➤ Why faster code generation isn’t producing faster software delivery
➤ What prevents AI agents from reaching production
➤ Why choosing the right AI model is only a small part of the challenge
➤ Why governance and guardrails cannot be added afterwards
➤ How AgentOps extends DevOps from code changes to autonomous decisions
➤ When deterministic automation is better than AI
➤ Why AI agents need identities, permissions and escalation paths
➤ How observability prevents autonomous systems from becoming dangerous black boxes
➤ What engineering teams need to operate AI safely at scale
➤ How Harness approaches everything that happens after code
Tell us what you loved about this Episode.
Website: itracusa.com
Contact: Feroz Khan
🎙️ This episode is brought to you by ITRAC.
ITRAC builds AI-ready, low-code-enabled teams.
We help organizations move from AI pilots to real production execution through:
• Strategic Staffing
• Enterprise AI & Low-Code Training
• Agentic AI Bootcamps
• CoE & Governance Advisory
AI creates momentum.
Execution creates advantage.
Whether you're scaling delivery or building your AI capability — we help you build it right.
