Databricks has raised the bar for enterprise AI assistants with the launch of Genie One — a “data-smart AI coworker” that connects to a company’s entire data ecosystem and achieves 84.5% accuracy on first attempt for real-world enterprise data analysis tasks. The announcement was made on the opening day of the Data + AI Summit 2026 in San Francisco, drawing more than 30,000 attendees to Moscone Center.
Genie One is designed to help business teams in finance, marketing, and sales move seamlessly from data insight to concrete action — a capability gap that has historically limited the practical value of business intelligence tools. The platform connects to a company’s complete data estate through native connectors and bidirectional integrations with everyday productivity tools including Gmail, Slack, and Microsoft Teams, making AI-powered data analysis accessible without requiring users to switch between applications.
The technical foundation of Genie One’s accuracy advantage is Genie Ontology, a context layer that automatically extracts knowledge from tables, queries, dashboards, pipelines, and connected applications, organizing it into what Databricks calls “a living graph of how a company works.” The system uses an authority-ranking approach similar to Google’s PageRank algorithm, evaluating the source of each definition, the authority of its author, usage frequency, and recency to determine which knowledge to prioritize in answering queries.
The accuracy numbers are compelling: on internal benchmarks using real-world enterprise data analysis tasks, Genie One answered 84.5% of questions correctly on the first attempt, compared to just 52.4% for the strongest general-purpose coding agent tested. Genie also delivered results twice as fast, addressing a critical bottleneck in enterprise data workflows where delays in getting answers can slow decision-making cycles.
Alongside Genie One, Databricks announced the evolution of Genie Spaces — of which customers have already created more than one million — into Genie Agents: domain-specific AI agents capable of taking autonomous multi-step actions, reasoning over both structured and unstructured data, and being created from a single natural language prompt. The company also introduced Genie App Builder for governed low-code application development and Genie ZeroOps for automated data quality and compliance at scale.
Foot Locker VP of Data and Analytics Matt Giunipero provided a real-world endorsement, stating that Genie Agents are “transforming how we lead” by providing executives with centralized AI-driven insights and “reshaping the way our business interacts with data.” With Databricks shifting to a pay-as-you-go pricing model and 150 free DBUs of LLM usage per month per user, Genie One is positioned to bring enterprise-grade AI data analysis within reach of organizations of all sizes.
