Databricks unveiled the updated Genie on Wednesday, saying it can handle extended, multi-step tasks without supervision. The agent uses an ontology built by Databricks to provide context from customers’ information on the platform.
In a demonstration using his own account, Ghodsi reviewed Genie’s analysis of weak areas in an employee satisfaction survey and its assessment of proposed fixes from executives. The agent had also summarized a sales pitch that went poorly, including the features the prospect said it needed before buying. It reviewed Ghodsi’s earlier podcast appearances to prepare likely questions, and suggested what he should say in the interview with the reporter.
Ghodsi said the tool can help fill a gap that grows as companies expand. Referring to Dunbar’s number—the theory that people can maintain stable relationships with about 150 others—he said a CEO at a company with 12,000 employees cannot know everyone. Employees may be less likely to bring a failed sales call or a potentially ineffective HR proposal directly to the chief executive.
“Anyone who has more than 150 employees, which is a lot of companies, a lot of people can get value out of this,” he said, speaking about AI agents and Genie. Databricks is pitching the product as an always-on agent for business leaders. Ghodsi also said he uses Genie personally, rather than alternatives including OpenClaw, Instinct or Muse.
He likened the potential role of agents in business to AI tools and software harnesses in programming. On concerns that agents could increase corporate AI costs, Ghodsi said Databricks’ Unity Gateway lets companies set usage limits and substitute less expensive open-weight models.
Comments (0)
No comments yet. Be the first!