Suffering from Success: Building Enterprise-Ready Internal Agents

We created IRIS to auto-triage PagerDuty incidents, but the real outcome wasn't just fixing one pipeline, it was unlocking a culture of rapid incident response across the business. We quickly realized our platform needed to be more than a single tool; it had to be a highly extensible foundation.

When n8n usage at TELUS exploded from a handful of small, discrete instances into thousands of workflows, our infrastructure didn't just grow, it forced us to rapidly re-architect our self-hosted GKE environment to sustain 4,700+ workflows. This session is our post-mortem: we'll share the lessons from that explosive growth phase, how we turned reliability from a struggle into our primary advantage, and how teams across TELUS are now cloning our patterns to build their own domain-specific agents.

Whether you're scaling from ten workflows to thousands, you'll leave with the strategies we used to turn a fragile experiment into a resilient, enterprise-grade automation engine.

• Understand how a single automation (IRIS) can unlock broader cultural change, in this case, faster incident response across the business
• Learn what forced TELUS to re-architect its self-hosted GKE environment when n8n usage exploded from a handful of instances to 4,700+ workflows
• Discover how the team turned reliability from a liability into their primary competitive advantage
• See how other teams across TELUS are now cloning these patterns to build their own domain-specific agents
• Identify strategies for scaling n8n infrastructure, whether you're going from ten workflows to thousands