Ready for the Agentic Era: Where Percona for MongoDB Is Going Next

For years, since 2015, Percona Software for MongoDB has meant one thing: running self-managed MongoDB with confidence. We brought in In-Memory Engine, LDAP, and OpenID Connect authentication, FIPS compliance, Audit logging, and File copy-based initial sync. Feature by feature, we turned the community server into an enterprise-ready database that hundreds of organizations trust, with no license fees and no lock-in.

Now, we have to brace for the new challenges. AI agents are moving from demos into production, and software architects are hitting the same wall: an agent is only as good as the context it can retrieve, rank, and retain. The smartest model in the world fails when the data layer serves stale, incomplete, or irrelevant results. In the agentic era, retrieval quality, and not model choice, decides whether an agent ships.

This keynote shows how Percona for MongoDB closes that gap. With full-text search, vector search, and hybrid retrieval natively integrated into an open, enterprise-ready platform, the same database that runs your operations becomes the memory and context engine for your agents. It has to be fast, accurate, and entirely under your control.

We’ll look at where Percona for MongoDB stands today, the real-world patterns it unlocks: From semantic product discovery to RAG and agent memory. Let’s explore where we’re heading: over the next 12 months, Percona will deliver a fully open, production-grade search and AI retrieval experience for MongoDB, so you can put agents in front of your customers without giving up freedom, flexibility, or control.

The agentic era needs an open data layer. Let’s build it together.