Databricks scales IT service with AI

As a rapidly growing software company at the forefront of AI and data analytics, Databricks needed a modern cloud IT service solution that would improve resolution times, establish stronger IT infrastructure, and scale easily and efficiently.

Previously, the company had used Zendesk and Spoke for ticketing but wanted a more cost-effective option with faster time to value and stronger automation capabilities. That’s when Databricks turned to Freshservice for IT service management (ITSM). Making the switch revamped the company’s overall IT service delivery, quickening ticket resolution times and, by implementing self-service, helped deflect support tickets that agents would have otherwise handled. Employee satisfaction hit 96%, driven by intuitive workflows and fast resolution times. Other departments soon took notice, and Databricks expanded Freshservice across HR, legal, security, and more. Databricks has since become one of the first Freshservice customers to adopt Freshworks' Model Context Protocol integration, connecting internal tools straight to service data for ticket analytics, pattern recognition, and automation at scale. Within the first month, 28% of new tickets originated through MCP.

Databricks views employee experience is a data problem as much as a service problem. Since consolidating onto Freshservice, our employees get faster answers and spend less time navigating between systems, so they can focus on their actual work instead of chasing tickets. That's the kind of experience Databricks wants every employee to have, no matter which team they're on.

The company

Databricks is an American enterprise software company with 10,000 employees that helps over 7,000 companies build and deploy data engineering workflows, machine learning models, analytics dashboards, and more. 

The challenge

Databricks needed to scale its IT service desk to improve efficiency and transparency for logging and resolving tickets. Initially using email and Slack to route IT requests from an employee base of nearly 5,000 at the time, the IT team faced longer-than-desired resolution times and struggled to track infrastructure issues and manage SLAs.

To address these challenges, Databricks sought a scalable, unified ITSM platform with AI capabilities to optimize IT investment decisions, deflect tickets, and help IT staff work more efficiently. But even as Databricks scaled Freshservice across departments, IT teams still lacked a way to apply AI directly against that growing volume of service data, as ticket analytics, pattern recognition, and bulk request handling remained manual, disconnected processes.

The solution

After evaluating several ITSM platforms, including ServiceNow, Zendesk, and SolarWinds, Databricks chose Freshservice for its no-code capabilities, automation features, and time to value. 

Using Freshservice's self-service knowledge base, employees can quickly find answers to common questions, reducing the tickets IT has to handle directly.

Other departments soon took notice, expanding Freshservice for Business teams across HR, legal, security, and learning and development. Employees now use a single self-service portal for all business functions, with tickets automatically routed to the right department's workspace under strict role-based permissions.

Yet as Freshservice scaled across the business, applying AI directly against that growing volume of service data remained a manual, disconnected process. To close that gap, Databricks turned to Freshworks' Model Context Protocol integration, becoming one of the first Freshservice customers to connect internal tools directly to Freshservice data. 

The rollout follows a deliberate access model: broad read-only access lets teams across the company query and analyze data safely, while read/write permissions remain limited to a smaller, trusted group. Use cases span ticket analytics and pattern recognition, bulk ticket creation through custom internal apps, and cross-referencing Freshservice data against other enterprise systems. 

Beyond ticket analytics, Databricks has connected roughly 48 tools through MCP, allowing IT to pull ticket populations on demand and catch trends and recurring issues before they show up on a dashboard. When tickets stall, backlog and SLA analysis surfaces exactly where and why. Managers use the same access to measure agent workload and ticket quality, turning what used to be manual review into real-time coaching. Knowledge base content gets audited the same way, so gaps and stale articles don't sit unnoticed. Even catalog requests can now fire automatically as part of larger workflows, and asset records stay current through the same reconciliation loop.

Engineers have also begun building on top of that access. Internal applications now handle bulk ticket creation, ticket rating, and fast individual-case lookups. Integrated with Freshservice's AI Agents, this same tooling rates and quality-scores tickets automatically, feeding into an AI-driven CSAT capability that evaluates ticket handling quality without manual review.

An additional layer of adoption has since emerged alongside that automation: engineers using AI coding assistants like Claude Code to query and act on Freshservice tickets directly, a pattern that has grown steadily since its introduction.

The approach reflects a deflection-first strategy. Databricks wants transactional, first-line work handled through self-service and automation, freeing IT to focus on complex issues that need hands-on expertise.

Impact

Using Freshservice in multiple departments improved efficiency and scalability organization-wide. Databricks was able to implement a clean data model foundation in Freshservice, leading to a meaningful improvement in deflection rate and employee experience. 

Consolidating from multiple point solutions to a single unified platform has also saved Databricks a significant amount of money. As the company maintains its rapid growth, Freshworks' unified ESM solution, now extended with AI and MCP, stands ready to power its next phase of scale.

And in the first month after rollout, 28% of new tickets created at Databricks originated through MCP, another step in Databricks' effort to make employee experience feel less like navigating tickets and more like getting an answer.