Databricks 详解 OpenAI 工具的三条企业落地路径
Databricks Details Three Enterprise Adoption Paths for OpenAI Tools
Databricks 于 2026 年 11 月 13 日发布企业采用指南,给出在受治理的公司数据上使用 OpenAI 工具的三种方式:通过 Genie 做 ChatGPT 数据分析、集中管控的 Codex 部署,以及以 Databricks Apps 托管的定制智能体。该方案延续 2025 年 9 月 25 日宣布、价值 1 亿美元的合作,OpenAI 提供 GPT 模型、ChatGPT、Codex 和 Agents SDK,Databricks 提供 lakehouse、访问规则与模型请求路由。三条路径共用 Unity Catalog 权限、Unity Gateway 成本控制和 Lakebase 存储,Unity Gateway 已正式可用,SQL、沙箱与网页搜索工具及 Unity Gateway 技能仍处于 beta。
正文
Databricks published an enterprise adoption overview on November 13, 2026, describing three ways to use OpenAI tools with governed company data: ChatGPT analytics through Genie, centrally controlled Codex deployments, and custom agents hosted as Databricks Apps.
The guide builds on a multi-year partnership announced on September 25, 2025. The original release valued it at $100 million and said native OpenAI models would be available to more than 20,000 Databricks customers across clouds. OpenAI co-founder and president Greg Brockman described AI without company data this way: “it’s like hiring someone with 20 PhDs and never inviting them to a meeting.”
Under the joint architecture, OpenAI supplies GPT models, ChatGPT, Codex and its Agents SDK. Databricks contributes the lakehouse, which stores governed company data, along with business definitions, access rules, model-request routing and application infrastructure. Genie Ontology adds company-specific context.
The business-user path targets employees who need figures but do not write SQL database queries. Through the Databricks Genie plugin, a ChatGPT user can ask a question in natural language; Genie identifies relevant data, applies approved metric definitions, runs the query in Databricks and returns graphs or charts. Genie Ontology records metrics, business definitions and relationships among company data, so a metric such as revenue follows the company’s own definition.
Access remains limited to data the user already has permission to see, and each answer links to its Databricks sources. ChatGPT Work can also recommend next steps and who should be involved, share findings through Slack or email, and carry out user-approved actions through connected tools.
OpenAI’s plugin listing also covers SQL generation, summaries of dashboards or Genie outputs, and packaging findings for teams. Availability depends on the plugin, subscription plan and workspace settings, and some connections require administrator setup or approval.
Unity Gateway Centralizes Codex Governance
For engineering teams, Codex sends requests through Unity Gateway instead of contacting OpenAI directly. Developers retain their existing workflow and authenticate with a Databricks credential, while the OpenAI API key stays in the gateway. Every request is logged, and engineering leaders can set team budgets, review spending by team and choose which GPT model handles each task.
The recommended setup installs the Unity Gateway command-line tool and runs ug codex. Databricks’ setup documentation also covers manual configuration and running Codex on Databricks compute. Eligible ChatGPT Go, Plus, Pro and Enterprise subscriptions can connect through an administrator-configured model-provider service, with requests billed against the ChatGPT subscription. Unity Gateway itself is generally available; the documented SQL, sandbox and web-search tools and Unity Gateway skills remain in beta.
The third path concerns custom agents for business processes that use data under strict security and governance rules. Databricks Model Serving runs OpenAI models beside that governed data, and no external API key is needed when Databricks serves the model. Teams can build and evaluate agents in Agent Bricks or use the OpenAI Agents SDK and deploy the result as a Databricks App.
The official app template supplies a conversational REST API, an interactive chat interface and MLflow evaluation code. Its MLflow AgentServer manages request routing, logging and error handling, with built-in tracing and observability. A ResponsesAgent interface provides compatibility with Databricks evaluation and deployment tools, while the Model Context Protocol connects the agent to tools and data sources.
Across these patterns, Unity Catalog supplies permissions, Unity Gateway handles request routing and cost controls, and Lakebase provides managed Postgres storage for agent memory and state. Databricks plans follow-up posts covering architecture, setup and customer lessons for each pattern.
That shared architecture gives business queries, coding requests and custom agents one permission and cost-control framework, making governance a common foundation as OpenAI use expands across a company.