NEWS · COMPANIES · #3430
Databricks updates Genie to run unsupervised multi-step, long-running business tasks
Databricks unveiled an updated Genie agent that can perform unsupervised, multi-step jobs spanning extended periods by using an ontology to surface context from data on the Databricks platform. CEO Ali Ghodsi demonstrated internal use cases—analyzing employee surveys, summarizing sales interactions and prepping interviews—and said Unity Gateway can limit usage and swap in cheaper models to control costs.
KEY POINTS
- Databricks unveiled an updated Genie agent that can perform unsupervised, multi-step jobs spanning extended periods by using an ontology to surface context from data on the Databricks platform.
- CEO Ali Ghodsi demonstrated internal use cases—analyzing employee surveys, summarizing sales interactions and prepping interviews—and said Unity Gateway can limit usage and swap in cheaper models to control costs.
- This update could accelerate adoption of always‑on AI agents in large organizations by enabling longer, multi‑step workflows that surface candid operational insights while offering cost controls.
WHY IT MATTERS
This update could accelerate adoption of always‑on AI agents in large organizations by enabling longer, multi‑step workflows that surface candid operational insights while offering cost controls.