Databricks Upgrades Genie AI Agent for Longer Business Tasks

Databricks said Wednesday it had upgraded Genie, its AI agent, to handle longer business assignments involving multiple steps. The company says customers can delegate a broad range of work for Genie to complete without ongoing supervision. The agent uses an ontology built by Databricks to draw context from information customers store on its platform.

CEO Ali Ghodsi has used Genie inside the company to examine an employee satisfaction survey, focusing on its lowest-scoring areas. He also asked the agent to assess possible responses suggested by executives and ruled out some fixes. In another example, Genie summarized a sales pitch that went poorly, including the features the prospective customer said it needed before buying.

The agent also helped Ghodsi prepare for a podcast interview by reviewing his previous appearances and suggesting questions he might face. He showed his recent Genie chats and charts using his own agent, rather than a staged product demonstration.

Ghodsi connected these uses to the challenge of managing a company with 12,000 employees. He put the potential value of AI agents in reach for companies with more than 150 employees. “Anyone who has more than 150 employees ... can get value out of this,” he said, referring to AI agents broadly and Genie in particular.

Databricks’ Row Zero acquisition was tied to a spreadsheet-style interface for Genie. The planned integration would let users work with data through natural-language prompts and a spreadsheet format.

Ghodsi also pointed to Databricks’ Unity Gateway as a way for companies to manage agent costs. The product lets customers cap usage and select less expensive open-weight models.