Databricks Genie One tackles AI context problem with Ontology layer

Databricks targets enterprise AI's accuracy gap

Databricks has introduced Genie One, an agentic AI assistant designed to automate business workflows by solving what the company identifies as AI's fundamental context problem. Unlike conversational analytics tools that retrieve information, Genie One executes tasks on behalf of workers by reasoning over both structured and unstructured data across an organization.

The company unveiled the platform at its Data + AI Summit in San Francisco, according to SiliconANGLE, which first reported the details. Chief Executive Ali Ghodsi framed the launch as a response to widespread disappointment with existing AI copilots, which he said have failed to deliver on early promises outside software engineering teams.

How Genie Ontology addresses fragmented business knowledge

The core innovation is Genie Ontology, a self-improving context layer that continuously scans and maps an organization's knowledge base. This includes data within Databricks, connected workplace applications, files, support tickets, chat platforms, and meeting transcripts. By extracting and maintaining this "ground truth," Genie One avoids the hallucination problem that plagues AI systems forced to guess when critical context is missing.

Ghodsi emphasized the stakes for regulated industries: "Most enterprise AI today is just guessing with false confidence, but that is not good enough for business. If you're a CFO and AI can't tell you why margins have changed, or you're a sales leader and it can't find your next upsell, that's not an AI problem, it's a context problem."

The assistant retrieves necessary context through SQL queries rather than attempting to reason from incomplete documents. Beyond answering questions, it provides interactive visualizations, alert capabilities, and integration with the Model Context Protocol to execute actions within third-party business tools.

New capabilities for data teams and business users

Databricks also released Genie Agents, which allow users to convert conversations with the original Genie chatbot into reusable automation workflows that inherit source data, instructions, and behavior patterns. A new Genie App Builder offers a low-code environment where business workers can upload context and generate data-connected applications secured by Databricks Unity Catalog.

For data engineering teams, the company updated Genie Code with project tracking and step-by-step review capabilities. Genie ZeroOps, a new background agent, autonomously monitors data pipelines, tables, and machine learning models to identify issues and propose fixes.

Databricks is pricing Genie One on a pay-as-you-go token consumption model rather than traditional SaaS licensing.

Why it matters

The gap between AI's technical capabilities and its practical business value often comes down to context. While large language models excel at pattern recognition, they struggle when the information needed to make reliable decisions is scattered across dozens of systems or exists only in institutional knowledge. If Genie Ontology can genuinely maintain accurate, up-to-date mappings of complex business environments, it could address one of the primary reasons AI automation projects fail in non-technical departments. The approach also highlights a broader industry shift toward grounding AI systems in verifiable enterprise data rather than relying on probabilistic reasoning alone.

Details of the announcement were first reported by SiliconANGLE.