Pinecone announced an integration between its Nexus knowledge engine and Microsoft (NASDAQ: MSFT) OneLake at Microsoft Build on June 3. The integration connects AI agents directly to enterprise data stored in Microsoft's unified data platform.

Pinecone Nexus creates structured, task-specific data contexts called artifacts from information already stored in OneLake. These artifacts allow AI agents to access relevant data without requiring manual imports or separate data pipeline management. The system applies role-based access controls and maintains data governance standards.

According to the company's press release, early testing showed a 95% reduction in frontier large language model token usage, 30x faster task execution, and completion rates above 90% compared to traditional data retrieval methods.

The integration uses KnowQL, Pinecone's query language designed for knowledge retrieval. AI agents specify their information requirements, output format, citation needs, and response time parameters through KnowQL queries.

"The data enterprises need to power their AI agents already live in Microsoft OneLake," said Ash Ashutosh, CEO of Pinecone. "Nexus builds task-specific artifacts from this data, and gives AI agents a clean, structured, cited interface through KnowQL, 30x+ faster and at a fraction of what traditional retrieval approaches cost."

Dipti Borkar, VP and GM of Microsoft OneLake and Fabric Ecosystem, said the integration reduces tool calls and token consumption while providing faster responses for enterprise customers.

The system maintains compliance requirements by assembling artifacts per task, applying permission controls, and providing source citations for all responses. Personal identifiable information is tagged during data ingestion and governed centrally.

Early access to Pinecone Nexus with OneLake integration is currently available. Pinecone serves more than 9,000 customers and 800,000 developers worldwide.