Inngest enables engineering teams to build reliable workflows. Recently we completed the rollout of new critical long-lived WebSockets infrastructure, enabling our customers to build low-latency agentic AI experiences. Together with AgentKit, a new SDK for AI agents, we have created the end-to-end infrastructure to help teams focus on creating tangible business value, instead of worrying about reliability.

Whether it’s equipping LLMs with tools to autonomously work on large codebases, building research assistants to search through academic databases and news archives, or creating personalized educational experiences, AI agents are here to stay, and they require robust infrastructure. In particular, we will focus on reliable, low-latency execution, observability, and scalability.

Key takeaways of this session are getting an understanding of the unique engineering challenges to enable AI agents, requirements for a robust WebSockets experience, as well as the significance of streamlining operations for improving team velocity while maintaining high reliability using cloud-native tooling for CD.