About the Product Demo
AI agents can react to events happening live, but only with access to the right data at the right time. That takes two things: grounding every answer in live operational data, and running the agent alongside the rest of your infrastructure. This can be accomplished with Retrieval Augmented Generation (RAG) and Aiven’s new Agents Runtime, and it’s what this session is about.
Most data architectures were built for people reading dashboards, not autonomous agents calling tools and making decisions. Bolt AI onto that same infrastructure and you get high token costs, slow responses, and answers nobody quite trusts.
We’ll show a working retail customer support agent, running on Aiven Runtime: querying PostgreSQL for order records, reading Kafka for real-time delivery events. Then we add DataHub, an open-source data catalog. The results are immediate. The agent now knows what data should exist, where it comes from, and why it might be missing. Instead of scanning blindly across every topic and table, it asks precise, targeted questions.
That’s what a unified platform changes. On Aiven, the open-source tools your agents depend on run under one control plane, across any cloud. Adding a new capability means provisioning a service, not building a whole new CI/CD pipeline. Every component is genuine open source: no proprietary forks, no vendor lock-in.
