Today, we’re thrilled to announce an exciting new step in our partnership with Databricks: the launch of the Box MCP server for Databricks. This integration unlocks immense value for organizations by seamlessly bridging the gap between unstructured content in Box and structured data in Databricks.
By bringing these two environments together, teams can now securely leverage their entire data estate right from their primary workflows without having to worry about complex data duplication, governance, or security risks.
Key takeaways:
- The integration connects unstructured files in Box with structured analytical data in Databricks without complex data duplication
- Users can securely access, cross-reference, and analyze Box content directly within Databricks Genie and the AI Playground
- The integration is governed by Unity Catalog and available through the Databricks Marketplace
The power of connecting structured and unstructured data
For data analysts, business analysts, data scientists, and data engineers, workflows almost always start with data. While Databricks serves as the center of gravity for running advanced analytics on structured metrics, critical context often lives across unstructured files like contracts, reports, and documentation stored in Box.
Databricks’ Model Context Protocol (MCP) support — available through Databricks Marketplace and governed by Unity Catalog — allows users to connect directly to third-party tools. By introducing the Box MCP server into Databricks’ partner ecosystem, users can now pull unstructured content from Box into their Databricks workflows to extract richer insights and get work done faster.
Bringing value across industries
Because almost every industry grapples with data silos between analytical platforms and operational content, the combination of Databricks and Box unlocks transformative use cases across multiple sectors, including:
- Consumer goods: Product teams tracking customer satisfaction ratings and sales volumes in Databricks can instantly ground their trends against focus group transcripts, user feedback forms, and product design briefs securely stored in Box
- Retail: Supply chain analysts monitoring inventory levels and real-time vendor fulfillment metrics within Databricks can instantly surface regional supplier agreements and SLA documents in Box to negotiate adjustments on the fly
- Public sector: Agency operations assessing program performance metrics or budget expenditures can instantly cross-reference regulatory policy frameworks, grant proposals, and compliance guidelines stored in Box to ensure adherence to strict oversight mandates
Seamless procurement and vendor workflows
To see this in action, imagine a company’s procurement team assessing whether or not to renew upcoming vendor agreements.
- Cross-reference data: The analyst asks Genie to look into Box for contracts expiring in the next six months and instantly cross-reference them with performance metrics in Databricks
- Identify risk and draft notices: Upon seeing that a vendor has underperformed, the analyst instructs Genie to automatically draft a termination notice based on the vendor’s contract clauses
- Automate collaboration: Instead of switching apps, the Box MCP server allows Genie to instantly upload the draft back to Box and invite the legal team as editors for review
By combining structured vendor metrics with unstructured contract information on a single platform, teams can make informed, rapid decisions without switching contexts.
Get started today
The Box MCP server is built to fit seamlessly into your existing data workflows. Users can find the Box MCP server in the Databricks Marketplace. Once connected, the Box MCP server will be readily available across all Genie products as well as in the AI Playground.
This launch marks just the first step in our growing partnership with Databricks as we continue to build a foundation for deeper data integrations. Explore the Databricks Marketplace today to try the Box MCP server and start maximizing the value of your data estate. To learn more about how to leverage the Box MCP server for Databricks, check out oursupport documentation.
