Pfizer's SRE teams faced rising complexity, constant interruptions, and high reliability expectations across research, manufacturing, and enterprise systems. Mission 40 tackled this head-on, aiming to cut operational workload by 40% across multiple teams while also migrating away from a costly vendor low-code platform to in-house automation. With 18 years of experience, Fuat Müminoğlu takes you inside the strategy, mindset, and execution that made this possible in just one year. Learn how careful prioritisation, multi-team collaboration, and targeted automation turned a massive challenge into clear, measurable impact, including the creation of a Continuous Improvement engine that scaled Mission 40's practices across the organization.
SRE is often introduced as a standalone team, yet reliability gains stall when it isn't embedded into everyday operations. In this session, Jay Rudrachaar explores how organisations can build SRE into the fabric of their business. Drawing on his work in global operations and AI driven observability at TIAA, Jay shows how the right operating model, shared ownership, and AI assisted tooling can shift teams from reactive incident response to prediction and prevention.
As AI tools spread across the enterprise, ITV's platform engineering team faced a familiar challenge: high potential value, but real risk around visibility, governance, and liability. Tom has been leading the build of ITV's new AI Agent Hub - an MVP now in controlled testing - designed to give teams a single, ITV-tailored place to access approved assistants, accelerate day-to-day work, and reduce "dark AI" usage. Powered by the platform foundations ITV has laid across Kubernetes, shared templates, and OpenTelemetry, the hub aims to make AI adoption traceable by default, so the organisation can confidently scale usage while protecting content, users, and the business.
Most organisations adopt SRE principles without adopting the "classic" model - and that's often the right call. The challenge is avoiding confusion, bottlenecks, and "SRE as the dumping ground" by agreeing what SRE is there to enable.
As systems become more distributed and change happens faster, "more dashboards" doesn't equal better reliability. The real shift is organisational: who owns signals, how teams respond, and how observability drives prioritisation rather than noise.
AI agents and LLMs are moving fast from experiments into everyday engineering, but most platforms were never built to support them at scale. Vincent Morel and Julien Tamisier take an inside look at how engineering systems must evolve as Schneider Electric builds an AI Platform for internal functions and business line use cases, while increasingly applying the same patterns and tools to develop the platform itself. From standardisation and developer workflows to AI-specific observability and session insights, this session shows how Schneider Electric is reshaping its platform so AI works reliably, visibly, and usefully across teams.
• Evolving engineering platforms to support AI workloads by embedding standards, guardrails, and observability by default
• Observing LLMs and agents through specialised telemetry, session data, and AI-focused SLI/SLOs
• Enabling developers to use AI confidently in everyday workflows without losing visibility or control, while feeding those learnings back into the platform itself
As Barclays expands the use of AI agents across its internal platforms, observability is becoming the foundation that turns experimentation into safe, scalable capability. By embedding observability into the architecture from day one and working in lockstep with security and engineering governance, Andy's team can monitor how agents behave, what data they touch, and where usability gaps emerge in real time. This deeper visibility is enabling the bank to boost productivity and respond to issues faster in a rapidly evolving AI landscape, all while maintaining compliance.