Databricks suggests key criteria for Genie Agent pilots

Databricks published a guide for choosing which Genie Agents—AI assistants that automate analytics workflows—deserve to be your team's first projects. Their advice, based on reviewing many failed and successful deployments, is to avoid agents with either overly broad or irrelevant use cases. Strong pilot candidates are high-impact, address frequent questions, rely on well-organized data, have a clear scope, and fit governance requirements. An executive sponsor is also critical for adoption.

• More than 1 million Genie Agents built in 2026
• Broad or niche agent pilots often see low adoption
• Best candidates automate frequent, high-impact decisions
• Well-governed data improves the agent's usefulness
• An internal sponsor is vital for adoption