As modern software systems become increasingly complex, Site Reliability Engineering (SRE) teams face mounting operational challenges. Traditional methods—manual triage, rule-based alerts, fragmented dashboards—are no longer sufficient to manage the scale and dynamism of today's distributed architectures. In response, the "AI-driven SRE" paradigm is emerging: agentic systems that emulate expert diagnostic reasoning, integrate service knowledge, and maintain a continuous improvement loop across the incident management lifecycle. Unlike legacy automation or general-purpose AI solutions, these SRE agents are tailored for the investigative, causal-analysis workflows unique to SRE—incorporating telemetry, change history, service topology, and post-incident knowledge. In their most advanced form, they transition from advising to semi-autonomous or fully autonomous operation, triaging incidents, hypothesizing root causes, executing remediation actions, and learning from outcomes in production.
In this talk we will walkthrough recent research and industry initiatives that embody this shift, highlight key architectural patterns (multi-agent coordination, model-context protocols, observability pipelines), and discuss avenues for safe adoption—governance, auditability, and gradual autonomy. In doing so, we argue that AI-driven SRE is poised not merely to augment reliability engineers but to transform how reliability is engineered and operated at scale.
João Freitas is General Manager and Engineering Lead for AI at Pager Duty. João leads PagerDuty AI initiatives and he is the main representative of the Lisbon office, being responsible for its growth, expansion, and culture. With about 20 years of experience in Software Development, Machine Learning, and as a People Manager, he was previously CTO at a startup in the area of Artificial Intelligence and has taken several roles at Microsoft in the areas of Speech Technologies and Data Engineering. João also holds a PhD in the areas of speech technology and human-computer interaction, filed several patents, and published over 40 articles in peer-reviewed international conferences and journals. He is also a regular speaker at AI conferences and the author and co-author of book chapters and one book.
