Databricks Says Internal Marketing Team Uses Data 3x More With Genie-Based Assistant Marge
Databricks details how its marketing organization tripled data use in decisions after deploying Marge, an internal AI analytics assistant built on Genie Agents.
Quick answer
What results does Databricks report from its internal Genie-powered marketing analytics assistant, Marge?
Databricks says its internal marketing team uses data three times more often in decisions after adopting Marge, a Genie Agents-based assistant built on a governed Marketing Lakehouse. The company reports over 85% adoption among marketers, 800-plus monthly questions answered, and a 25% drop in flagged incorrect responses.
Key takeaways
- Databricks says its marketing team now uses data three times more often in decision-making after deploying Marge, an internal assistant built on Genie Agents.
- The company reports that more than 85% of its marketing organization uses Marge, with usage growing 50% quarter over quarter and more than 800 questions answered monthly.
- Databricks says flagged incorrect responses fell 25% after it added verified answer sets, example queries, and documented business definitions in Unity Catalog.
- The company credits its rollout strategy, starting with a single use case (email campaign performance) and expanding based on user feedback, for driving adoption.
- Databricks says one BI manager spends about one hour per week maintaining Marge, which the company says allowed it to scale access to insights without adding analytics headcount.
What Databricks announced
Databricks published a company blog post detailing how its own marketing organization uses an internal AI assistant called Marge, which the company describes as its marketing implementation of Genie Agents, its conversational analytics technology. The post was written by Elizabeth Dobbs, Databricks' AVP of Marketing Technology, along with Thomas Russell, Katy Yuan and Sydney Sundell.
According to Databricks, Marge lets marketers ask questions in natural language and receive answers grounded in a "Marketing Lakehouse," a governed data environment the company built to unify campaign, web analytics, CRM, event and sales data behind consistent definitions and metrics.
Reported usage and results
Databricks says the tool has produced the following outcomes inside its marketing department:
- Marketers use data three times more often when making decisions.
- More than 85% of the marketing organization uses Marge.
- Usage has grown 50% quarter over quarter.
- Marge handles more than 800 questions per month and has answered more than 5,000 questions in total.
- Flagged incorrect responses have decreased by 25% as the system has been refined.
- Databricks says it scaled access to insights without adding analytics headcount.
The company says marketers can ask Marge questions such as how an email campaign performed, which programs influenced pipeline in a given quarter, or where to adjust campaign spend, and that Marge translates those questions into queries run against governed enterprise data. Access runs through Genie One, which Databricks describes as an "AI cowork experience for business users" that routes requests to specialized Genie Agents covering areas such as web performance, digital analytics and marketing planning. For more complex requests, the company says an "Agent mode" can evaluate multiple steps to produce deeper analysis.
Unity Catalog, Databricks' governance product, provides centralized metadata, lineage and role-based access control so that each user sees only data they are authorized to view, according to the post.
How Databricks says it built trust into the system
Databricks outlines four practices it credits for making Marge more accurate and reliable:
- Documenting data and table relationships in Unity Catalog, including business-specific definitions such as what counts as a marketing-qualified lead.
- Encoding verified answers and example question-and-query pairs, reviewed by domain experts, for high-value or complex questions.
- Teaching Marge business terminology and instructing it to ask clarifying questions when a request is missing details such as time period or region.
- Running continuous feedback and evaluation loops, including user ratings and benchmark questions checked against known answers.
The company says one BI manager spends about one hour per week reviewing feedback and maintaining Marge, an investment it credits with the 25% drop in flagged incorrect answers.
Rollout approach
Databricks says it drove adoption by working directly with marketers to identify their most common questions and the language they use, then starting with a single, narrow use case, email campaign performance, before expanding scope. The company says this let it validate accuracy and build user confidence before broadening the assistant's coverage.
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?
- Databricks describes Marge as an internal AI analytics assistant built on its Genie Agents technology, used by its own marketing team to answer natural-language questions against governed data.
- What results has Databricks reported from using Marge?
- The company says marketers now use data three times more often in decisions, more than 85% of its marketing organization uses Marge, and flagged incorrect answers have dropped 25%.
- How does Marge access marketing data?
- Databricks says Marge draws on a governed "Marketing Lakehouse" and uses its Unity Catalog product for metadata, lineage and role-based access control.
- How did Databricks roll out Marge to its marketing team?
- The company says it began with a single use case, email campaign performance, before expanding scope based on user feedback and adoption.