Introducing Genie One, Genie Agents, and Genie Ontology

Despite the progress in LLMs and consumer chat agents, most enterprise teams still struggle to put AI to work on real business questions. The reason is that getting insights out of data - the ground truth of the business - is hard even with the latest generations of models and agents.尽管LLM和消费级聊天代理取得了进展，但大多数企业团队仍然难以将AI应用于真实的业务问题。原因在于，即使是使用最新一代的模型和代理，从数据（业务的真实依据）中获取洞察仍然很困难。

The reason for this is that the business context required to use data is scattered across dashboards, queries, pipelines, wikis, tickets, documents, and chat threads. When AI doesn’t easily find the information it needs, it fills in the gaps with inference, producing answers that are generic at best and wrong at worst. The current generation of agents often go through a process of iterative probing that is extremely slow and costly, and forcing a compromise in quality. This has resulted in unacceptably poor performance for truly data-driven decision making and actions.其原因是使用数据所需的业务上下文分散在仪表盘、查询、管道、维基、工单、文档和聊天线程中。当AI无法轻松找到所需信息时，它会用推理来填补空白，产生的答案充其量是通用的，最坏情况下是错误的。当前一代的代理通常经历一个迭代探查的过程，极其缓慢且成本高昂，迫使质量妥协。这导致了真正数据驱动的决策和行动性能低得令人无法接受。

Today, we are announcing our solution to this problem:今天，我们宣布了解决这个问题的方法：

1. Genie One: a data-smart AI coworker that works across all of your data,Genie One：一个数据智能的AI同事，可在你的所有数据上工作，
2. Genie Agents: a way for everyone to create dedicated agents to automate work, andGenie Agents：一种让每个人都能创建专用代理来自动化工作的方式，以及
3. Genie Ontology: an automatic and secure context store that enables agents to achieve superior accuracy and performance.Genie Ontology：一个自动化的安全上下文存储，使代理能够实现卓越的准确性和性能。

Genie One: The data-smart AI coworker for every teamGenie One：面向每个团队的数据智能AI同事

Genie began as a conversational analytics assistant in Databricks AI/BI. Genie One is the next step: a data-smart AI coworker designed to help users move from insight to action.Genie最初是Databricks AI/BI中的一个对话分析助手。Genie One是下一步：一个数据智能的AI同事，旨在帮助用户从洞察走向行动。

• For all of your apps and data. Genie One connects insights from across your entire data estate through a rich ecosystem of native connectors. Through Lakehouse federation, Lakeflow Connect, and new two-way integrations to everyday business tools (e.g., Gmail, Slack, Teams), Genie One can extract insights and orchestrate actions with any system you may have.适用于你的所有应用和数据。Genie One通过丰富的原生连接器生态系统，连接来自整个数据资产的洞察。通过Lakehouse联邦、Lakeflow Connect以及与日常业务工具（例如Gmail、Slack、Teams）的新双向集成，Genie One可以提取洞察并协调与你可能拥有的任何系统的行动。
• Full co-work capabilities. Genie One provides a full suite of agentic-cowork capabilities, including schedules and alerts, monitoring, document creation, custom skills, and custom MCP support. Sales leaders can ask Genie to prepare a daily brief for all their customer meetings, combining the context from their calendar, email, and hard data from their Lakehouse and other connected systems. General managers can simply paste in last month’s business review document, and ask Genie to update it with the latest information from their inventory systems, along with the transcripts from their team discussions. The possibilities are endless.全面的协同工作能力。Genie One提供一整套代理协同工作能力，包括计划和警报、监控、文档创建、自定义技能和自定义MCP支持。销售主管可以要求Genie为其所有客户会议准备每日简报，结合来自日历、电子邮件以及Lakehouse和其他连接系统的硬数据。总经理只需粘贴上个月的业务审查文档，要求Genie用来自库存系统的最新信息以及团队讨论的记录进行更新。可能性是无限的。

Because enterprise work happens across a full stack of tools and surfaces, we’re bringing Genie to everywhere work happens—starting by embedding it natively into Slack and Microsoft Teams. Users can simply @mention Genie in any conversation to ask questions in natural language and get accurate answers in seconds. Genie can also be used in public channels and threads, helping teams collaborate without switching context. Every response is governed, secure, and tailored to what each user is authorized to access.由于企业工作发生在全套工具和界面上，我们将Genie带到工作发生的任何地方——首先将其原生嵌入Slack和Microsoft Teams。用户只需在任何对话中@提及Genie，即可用自然语言提问，并在几秒钟内获得准确答案。Genie还可以在公共频道和线程中使用，帮助团队无需切换上下文即可协作。每个响应都经过监管、安全，并根据每个用户被授权访问的内容进行定制。

For users on the go, we’re launching new iOS and Android apps that put Genie in your pocket. Users can ask questions, get alerts, and take action on insights grounded in your company data, from anywhere.对于移动用户，我们推出了新的iOS和Android应用，将Genie放入你的口袋。用户可以从任何地方提问、获取警报并对基于公司数据的洞察采取行动。

And for organizations who have adopted an existing AI agent or developed their own, we’re also announcing the Genie MCP App, which allows those users to benefit from Genie without having to change their workflows.对于已采用现有AI代理或自行开发的组织，我们还宣布了Genie MCP应用，允许这些用户无需更改工作流程即可受益于Genie。

Our investment in building the Uplight Data Platform on top of Databricks is paying off in powerful new ways. By bringing Genie One capabilities to our data, we’re enabling teams across Uplight to explore, discover, and innovate with more speed, confidence, and creativity than ever before. This is the promise of data democratization - enabling a culture where curiosity, data-informed decision-making, and innovation can happen at every level of the company.—Micaela Christopher, Director of Data Science and Engineering, Uplight我们在Databricks之上构建Uplight数据平台的投资正以强大的新方式得到回报。通过将Genie One功能引入我们的数据，我们使Uplight的各个团队能够以前所未有的速度、信心和创造力进行探索、发现和创新。这就是数据民主化的承诺——在公司各个层面培养好奇心、数据驱动决策和创新的文化。——Micaela Christopher，Uplight数据科学与工程总监

Genie Agents: from a single prompt to an autonomous agentGenie Agents：从单个提示到自主代理

Databricks customers have created more than a million Genie Spaces—curated, governed chat experiences scoped to specific topics, with verified logic and benchmarks. Now, Genie Spaces is evolving into Genie Agents: curated, domain-specific AI agents that:Databricks客户已创建超过一百万个Genie Spaces——针对特定主题策划、受监管的聊天体验，具有经过验证的逻辑和基准。现在，Genie Spaces正在演变为Genie Agents：策划的、特定领域的AI代理，它们：

• Take autonomous action. Genie Agents use the same tools Genie One does—MCP connections, scheduled tasks, document and artifact generation, and writes to external systems—to complete multi-step workflows without need for oversight or intervention.采取自主行动。Genie Agents使用与Genie One相同的工具——MCP连接、计划任务、文档和工件生成以及对外部系统的写入——来完成多步骤工作流，无需监督或干预。
• Reason over unstructured data, not just tables and views. Agents can be grounded in documents, files, and knowledge sources alongside structured data, covering the full context of a real business problem.不仅对表格和视图，还能对非结构化数据进行推理。代理可以基于文档、文件和知识源以及结构化数据，覆盖真实业务问题的完整上下文。

Best of all, creating an agent is as simple as describing what you want: spin up a Genie Agent from a single prompt in Genie One or Genie Code, scope it, benchmark it, and share it for teammates to use or customize. Genie Agents let domain experts scale their expertise by turning trusted rules, data, and workflows into coworkers the whole team can rely on.最棒的是，创建一个代理就像描述你想要的一样简单：在Genie One或Genie Code中从一个提示启动Genie Agent，设定范围，进行基准测试，然后分享给队友使用或自定义。Genie Agents让领域专家通过将受信任的规则、数据和工作流转化为整个团队可以依赖的同事来扩展他们的专业知识。

At Foot Locker, Genie Agents are transforming how we lead. They provide our executives and business teams with a centralized space to harness AI-driven insights across every North American banner we operate. As we scale Genie to the enterprise, it's reshaping the way our business interacts with data and makes the decisions that matter most. Genie isn't just a tool; it's the engine driving self-service insights across our organization—Matt Giunipero, VP of Data & Analytics and Krish Lakshminarayanan, VP, AI, Data & Analytics, Enterprise Architecture, Foot Locker在Foot Locker，Genie Agents正在改变我们的领导方式。它们为我们的高管和业务团队提供了一个集中空间，以利用AI驱动的洞察，涵盖我们运营的每个北美品牌。随着我们将Genie扩展到企业，它正在重塑我们的业务与数据交互以及做出最重要决策的方式。Genie不仅仅是一个工具；它是驱动我们整个组织自助式洞察的引擎。——Matt Giunipero，数据与分析副总裁；Krish Lakshminarayanan，AI、数据与分析、企业架构副总裁，Foot Locker

Genie Ontology: a living context graph for your businessGenie Ontology：业务的生活上下文图谱

Genie One and Genie Agents are powered by Genie Ontology, an automatic context layer. Genie Ontology automatically extracts snippets of knowledge from tables, queries, dashboards, pipelines, and connected apps, and organizes that knowledge into a living graph of how a company works and what the data inside actually means. Genie has context about where to look, what to trust, and how to answer in a way that reflects how the company actually uses its data. That includes metric definitions, business terms, unique calculations, and the relationships between concepts, metrics, tables, and teams.Genie One和Genie Agents由Genie Ontology驱动，这是一个自动化的上下文层。Genie Ontology自动从表、查询、仪表盘、管道和连接的应用中提取知识片段，并将这些知识组织成一个生活图谱，反映公司如何运作以及数据内部的真正含义。Genie拥有关于在哪里查找、信任什么以及如何以反映公司实际使用数据的方式回答的上下文。这包括指标定义、业务术语、独特计算以及概念、指标、表和团队之间的关系。

One key innovation of Genie Ontology is its approach to determining authority. Using an approach similar to PageRank, Genie Ontology weighs where a definition came from, the relative authority of that source’s author, how often people rely on it, how closely it ties to certified and widely-used assets, and how fresh it is. Then, Genie answers from the sources that carry the most weight. It also enforces the permissions of each source by only showing you the content that you actually have permissions to see. The result is that Genie solves the context problem, without asking your teams to hand-curate it or manage a separate permissions system.Genie Ontology的一个关键创新是其确定权威性的方法。使用类似于PageRank的方法，Genie Ontology权衡定义的来源、来源作者的相对权威性、人们依赖它的频率、它与经过认证和广泛使用的资产的紧密程度以及它的新鲜度。然后，Genie从权重最高的来源回答。它还通过仅显示你实际有权限查看的内容来强制执行每个来源的权限。结果是，Genie解决了上下文问题，而无需你的团队手工策划或管理单独的权限系统。

Our internal benchmark of real-world enterprise data analysis tasks have shown that Genie Ontology significantly improves agent performance on complex, enterprise data questions. In our testing, Genie answered 84.5% of questions correctly on the first attempt, while the strongest general-purpose coding agent managed just 52.4% — and the weakest only 25%. And Genie doesn’t trade off accuracy for speed. Genie delivers high accuracy and low latency, 2× faster than the strongest coding agent. 我们对真实企业数据分析任务的内部基准测试表明，Genie Ontology显著提高了代理在复杂企业数据问题上的性能。在我们的测试中，Genie首次尝试的正确回答率为84.5%，而最强通用编码代理仅为52.4%，最弱的只有25%。而且Genie不会以牺牲准确性换取速度。Genie提供高准确性和低延迟，速度是最强编码代理的2倍。

Governed and secure by design设计与监管和安全

To roll out any AI tool across a company, leaders and IT teams need confidence that it’s governed, secure, and ready to scale. That’s why Genie One includes a full suite of admin governance tools designed to help organizations deploy AI across their teams.要在公司内部推广任何AI工具，领导者和IT团队需要确信它受到监管、安全且可扩展。这就是为什么Genie One包含一整套管理治理工具，旨在帮助组织在其团队中部署AI。

Like every Databricks product, governance and security sit at the heart of Genie. Permissions are enforced by default on every answer through source-native ACLs or Unity Catalog. MCP, tools and costs are governed by the Unity AI Gateway, providing a single pane of governance for admins.与所有Databricks产品一样，治理和安全是Genie的核心。每个答案默认通过源原生ACL或Unity Catalog强制执行权限。MCP、工具和成本由Unity AI Gateway管理，为管理员提供单一治理面板。

Getting started开始使用

Genie One is the data-smart AI coworker every business user needs: it understands enterprise context, works across the tools where work happens, and is governed by design.Genie One是每个业务用户所需的数据智能AI同事：它理解企业上下文，跨工作发生的工具工作，并且按设计受到监管。

To try Genie One, see our documentation, install the mobile app for iOS or Android, or contact your Databricks account team.要试用Genie One，请参阅我们的文档，安装iOS或Android移动应用，或联系你的Databricks客户团队。