Tech Transformed podcast

Why Do 1 in 4 AI Agents Run Without Monitoring?

0:00
27:36
Recuar 15 segundos
Avançar 15 segundos

Picture an AI agent left running overnight. It edits a config file, closes a ticket and moves on to the next task. Every dashboard stays green. No human sees why it acted.

That is the reality for roughly one in four organisations with agents in production.

New Relic's 2026 Observability Forecast surveyed 2,575 IT and engineering leaders. It found that 93 per cent of enterprises are adopting AI agents, yet 24.6 per cent run them unmonitored.

"We're in a space race," Jemiah Sius, Vice President of Market Strategy and Developer Relations at New Relic, told BARC's Kevin Petrie on the Tech Transformed podcast. Teams are setting aside enterprise software discipline to move fast and prove return on investment, he said. It takes only one major outage, he warned, for leaders to slow down.

In this episode of Tech Transformed, Kevin Petrie, VP of Research at BARC, talks to Jemiah Sius, Vice President of Market Strategy and Developer Relations at New Relic, about how teams should inspect what AI agents actually do, and how to scale past human sign-off safely.

They discuss why teams ship agents first and instrument later, the Hugging Face sandbox breach, why there is no control without observability, how to earn trust in autonomy by limiting an agent's tools and MCP servers, and how pre-deployment testing can stop a bad agent release before it reaches production.

Key takeaways

  • 93% of organisations adopt AI agents; 24.6% run them unmonitored
  • Teams ship agents first because leaders need to show a return
  • Observability has traditionally watched, not blocked
  • Limit an agent's tools and MCP servers before removing human approval
  • An orchestrator plus narrow sub-agents adds determinism
  • Pre-deployment testing can catch jailbreaks and PII leakage
  • Sius expects EU rules to demand proof of agent behaviour

Chapters

  • 00:00 The Rise of AI Agents and Observability Challenges
  • 02:23 Understanding the Gap in Monitoring AI Agents
  • 05:51 The Risks of Shadow AI and Governance Issues
  • 08:49 Building Trust in Autonomous AI Agents
  • 12:56 The Role of Observability in Security and Governance
  • 18:17 Real-World Use Cases of Observability in Action
  • 22:41 Future Trends in AI Agent Observability
  • 25:45 Strategies for CISOs and CIOs to Enhance Observability

Research mentioned in this episode New Relic 2026 Observability Forecast: https://newrelic.com/sites/default/files/2026-09/new-relic-2026-observability-forecast-report.pdf

BARC: AI is in production at 97% of organisations: https://barc.com/news/ai-is-in-production-at-97-of-organizations-but-no-control/

Further reading:

The Hugging Face sandbox breach: https://techwireasia.com/2026/07/openai-agent-sandbox-breach-hugging-face/

Agent observability on the Claude Platform: https://claude.com/resources/articles/agent-observability-on-the-claude-platform

EU AI Act: what it means for UK businesses: https://em360tech.com/tech-articles/eu-ai-act-what-it-means-uk-businesses-and-how-stay-compliant

Shadow AI and shadow agents: https://em360tech.com/podcasts/shadow-ai-shadow-agents-whats-actually-changed-2026

How to monitor AI agents in production: https://em360tech.com/podcasts/how-do-you-monitor-ai-agents-production-without-breaking-incident-response

Learn more about New Relic: https://newrelic.com

Subscribe to Tech Transformed for weekly conversations on how technology is changing the way businesses run.

Website: https://em360tech.com

LinkedIn: https://www.linkedin.com/company/em360tech

𝕏 X: https://x.com/EM360Tech

#AIAgents #AgentObservability #AIGovernance #AgenticAI #NewRelic #AISecurity #ShadowAI #OpenTelemetry #TechTransformed #EM360Tech

AI agents, agent observability, AI observability, agentic AI, AI governance, AI security, shadow AI, sandbox escape, MCP, OpenTelemetry, EU AI Act, human in the loop, New Relic, BARC, Tech Transformed

Mais episódios de "Tech Transformed"