Incident sheet · /enterprise-ai
- System
- Enterprise AI
- Filed
- Read
- 3 min · 619 words
- Severity
- Routine
Databricks Genie Now Runs Long Jobs — and Tells CEOs the Truth
Databricks updated Genie to run multi-step jobs unsupervised. CEO Ali Ghodsi uses it to surface what 12,000 employees won't tell him, citing Dunbar's 150.

Event log
- Databricks unveiled the updated Genie agent on Wednesday, capable of unsupervised multi-step jobs
- Genie draws context from customer data via an ontology built on the Databricks platform
- CEO Ali Ghodsi cites Dunbar's number of 150; Databricks has 12,000 employees
- Ghodsi says costs can be capped via Unity Gateway usage limits and open-weight models
- He positions Genie against security-questionable alternatives like OpenClaw, Instinct and Muse
Databricks shipped an updated version of its AI agent Genie on Wednesday, and the upgrade matters: the agent can now complete multi-step jobs that take extended work to finish, unsupervised.
The mechanism is an ontology Databricks built to feed Genie context from the information customers already store on the Databricks platform. Corporate leaders can assign the agent a wide range of tasks and let it run, the company says.
CEO Ali Ghodsi did not offer a canned product demo. He fired up his actual agent and walked through his most recent chats and charts.
What does Ghodsi actually use it for?
The use cases were unglamorous and, for that reason, credible. Genie analyzed the parts of an employee satisfaction survey where Databricks scored poorly. It then helped him rate — and discard — fixes for those low scores that some executives had proposed.
It summarized a large sales pitch to a prospect that went badly, including the features the prospect said it needed before buying. It prepped him for a podcast interview by reviewing his previous appearances and drafting likely questions. During the conversation itself, he even had it tell him what to say. Its advice, reportedly, was not bad.
The pattern is the interesting part. Ghodsi uses Genie to surface the things a chief executive no longer hears. It functions as a source of unvarnished truth about what is not going well and needs fixing.
Why 150 employees changes everything
Ghodsi has a theory about why this matters. "Have you heard of Dunbar's number? The magic number is 150," he said when the point came up.
The number is named for British anthropologist Robin Dunbar and marks the theoretical maximum of stable relationships a human can maintain. In a small startup, the CEO knows everyone, holds direct relationships, and people feel safe delivering negative information straight.
Databricks has 12,000 employees. Ghodsi admits he cannot know them all. At that size, a salesperson will not confess a disastrous call to the CEO. An HR manager will not admit their plan for raising employee satisfaction will probably fail.
"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 — of AI agents generally and of Genie specifically. That is a broad addressable market, stated without much subtlety.
What does this change for enterprise AI buyers?
Databricks is now pitching Genie as an always-on agent for business leaders, not a query tool. Ghodsi's own daily use — what the industry calls dogfooding, from "eating your own dog food" — anchors the positioning. He says he also runs it in his personal life, in preference to what he calls security-questionable alternatives such as OpenClaw, Instinct or Muse.
His larger argument: Genie can do for the business side what AI agents and harnesses are already doing for programmers. Harnesses, in this framing, are the software wrapped around an AI model that lets it use tools. The pitch is that executives get the same productivity lift developers have enjoyed.
What about the cost question?
Unsupervised, always-on agents invite an obvious worry: runaway spend. Ghodsi has an answer ready.
Costs can be controlled through Databricks' Unity Gateway product, he said. Unity Gateway lets companies set usage limits and swap in less expensive open-weight models when the budget demands it.
So the decision framework for buyers is fairly clean. If your organization sits past the Dunbar line — and most do — Databricks is betting the CEO will pay for an agent that reports what the org chart filters out. The unanswered question, as ever, is whether employees will trust a platform-owned agent with the truths they currently withhold.
via techcrunch.com (Source)
More from Tom Whitfield
Show full bio
Market editor covering people, launches and day-to-day practice in enterprise information technology and the leaders who run at CIO Monitor.
185 articles
The newsroom digest
One email a day with the stories that matter. No noise.
Cross-references
- Databricks adds always-on Genie agents and ontology
- Databricks Genie Goes GA in Teams and Plugs Into Microsoft Copilot
- Before You Sign: The Traceability Demands Data Leaders Should Put on Vendors
- OpenAI’s Business Hub makes ChatGPT more search-like
- AI Exposed the Consulting Billable Hour — CIOs Are Now Renegotiating