Databricks Says Internal AI Assistant Tripled Marketers' Use of Data
Databricks says its marketing team built Marge, a Genie Agents-based analytics assistant, that lifted data usage 3x and reached over 85% adoption.
Quick answer
What did Databricks announce about its Marge AI analytics assistant for marketing?
Databricks says it built Marge, an AI analytics assistant powered by Genie Agents and grounded in a governed Marketing Lakehouse, that lets marketers ask questions in natural language. The company reports marketers now use data three times more often in decisions, with adoption exceeding 85% of the marketing organization.
Key takeaways
- Databricks says it built Marge, a marketing-specific implementation of Genie Agents, on top of a governed Marketing Lakehouse that unifies campaign, web, CRM and sales data.
- Databricks reports marketers now use data three times more often in decisions, with adoption exceeding 85% of the marketing organization and usage growing 50% quarter over quarter.
- According to Databricks, Marge handles more than 800 questions per month and has answered over 5,000 total, while flagged incorrect responses have dropped 25% as the system improved.
- Databricks says it used Unity Catalog for centralized metadata, lineage and role-based access controls, and combined this with verified answers, example queries and business-language guidance to make Marge more accurate.
- Databricks says the rollout began with a single use case, email campaign performance, before expanding scope, and that maintaining Marge now takes one BI manager about one hour per week.
Databricks says its marketing organization built an internal AI analytics assistant called Marge that has tripled how often marketers use data to make decisions, according to a blog post from the company.
The assistant is a marketing-specific implementation of Genie Agents, Databricks' conversational analytics technology, and is grounded in what the company calls a governed Marketing Lakehouse. That lakehouse unifies data from campaign platforms, web analytics, CRM systems, event tools, advertising channels and sales data, which Databricks says previously lived in separate systems with inconsistent definitions.
What Databricks reported
According to the company, since deploying Marge:
- Marketers use data three times more often to make decisions.
- More than 85% of the marketing organization uses Marge.
- Usage has grown 50% quarter over quarter.
- Marge handles more than 800 questions each month and has answered over 5,000 in total.
- Flagged incorrect responses have decreased by 25% as the system improved.
- Databricks says it scaled access to insights without adding analytics headcount.
Databricks says marketers interact with Marge by asking questions in natural language, such as how an email campaign performed or which programs influenced pipeline in a given quarter. The assistant translates those questions into analytical queries and returns answers drawn from the governed data, according to the company. Databricks says marketers access Marge through Genie One, described as an AI cowork experience for business users that also includes dashboards and apps, and that Genie One routes requests to specialized agents built for domains like web performance, digital analytics and marketing planning.
How Databricks says it built accuracy and trust
The company describes four mechanisms it used to make Marge more accurate and reliable. First, Databricks says it used Unity Catalog to centralize metadata, table descriptions, lineage and access controls, with marketing stakeholders reviewing and enriching AI-generated descriptions with business context. Second, the company says it encoded verified answers and example queries for common, high-value questions, such as conversion rates and customer lifetime value, reviewed by domain experts. Third, Databricks says it gave Marge behavioral guidance for interpreting company-specific terminology, such as treating "spend" and "investment" as equivalent to a field named "cost," and instructed the assistant to ask clarifying questions when requests are missing information like time period or region. Fourth, the company says it built continuous feedback and evaluation loops, where users rate responses and a BI manager reviews the ratings, which Databricks says has helped reduce the rate of flagged incorrect answers by 25%. The company says this maintenance now takes about one hour per week.
Rollout approach
Databricks says it began by working directly with marketers to understand the questions they asked most often and the language they used, then started with a single focused use case: email campaign performance, using only the campaign, recipient and engagement data needed to answer those questions. The company says this narrow initial scope made it easier to validate accuracy and demonstrate value before expanding to additional data and use cases.
Source: Databricks Blog, "How Databricks' marketers use data 3x more with Genie, an AI analytics assistant," published September 15, 2026.
Frequently asked questions
- What is Marge?
- According to Databricks, Marge is a conversational analytics assistant built with Genie Agents, grounded in the company's governed Marketing Lakehouse, that lets marketers ask questions in plain English and get governed answers.
- What results does Databricks report from using Marge?
- Databricks says marketers now use data three times more often in decisions, adoption exceeds 85% of the marketing organization, and usage has grown 50% quarter over quarter.
- How does Databricks say it kept Marge's answers accurate?
- Databricks says it documented data relationships in Unity Catalog, encoded verified answers and example queries, taught Marge business-specific terminology, and built continuous feedback and evaluation loops.
- How did Databricks roll out Marge to marketers?
- Databricks says it started with users' real questions, began with a single focused use case (email campaign performance), and expanded scope from there.