Salesforce and Databricks have announced an expanded partnership designed to connect Salesforce Data Cloud, Agentforce and Databricks' Data Intelligence Platform, giving employees and AI agents access to the same enterprise data. The announcement was made at the Databricks Data + AI Summit on 16 June. Salesforce frames the move as a way to link enterprise data to the customer context, permission structures, approval chains and processes organisations need to act on it with confidence.
Many organisations have invested heavily in AI over the past two years, yet a large share of initiatives stall once they leave the pilot stage, often because the underlying data remains scattered across disconnected systems. This partnership reflects a theme now running through much of enterprise AI whereby scaling agents is increasingly a data problem rather than a model problem.
As Rahul Auradkar, President and GM, Data Foundations at Salesforce, observes: “The challenge is no longer building more agents. It’s giving agents the trusted data, business context, governance, and workflows required to operate safely at enterprise scale.”
What the partnership includes
The expansion builds on the existing Zero Copy relationship between Salesforce Data 360 and Databricks Unity Catalog, adding governance features including Federated Authentication, planned identity mapping, governance interoperability and metadata-aware access controls so that permissions don't need to be rebuilt separately in each platform.
A new Federated Search feature lets Agentforce agents pull information from Databricks, and lets Databricks users query Salesforce data, through one unified search experience. MuleSoft's Agent Scanners add visibility into Databricks activity within Agent Fabric, supporting governance at scale. Finally, new integrations also bring Databricks tools, such as Genie, directly into Slack, surfacing data insights and security alerts where employees are already working.
Why data has become enterprise AI's biggest challenge
AI agents are only as useful as the information they can reach, and for many businesses that information still sits in fragmented customer, operational and business systems that were never designed to talk to each other. The result is often inconsistent customer experiences and unreliable agent outputs, even when the underlying model is capable. For CX teams specifically, this fragmentation can prevent agents from delivering the accurate, personalised and context-aware interactions that customers now expect. The bottleneck for agentic AI has shifted from agent creation to data access and governance, a problem many CX organisations still get wrong.
What this means for CX leaders
For customer experience leaders, better data access could translate into more reliable agent performance, sharper personalisation and faster decision-making, particularly where human agents and AI systems need to collaborate on the same case or customer record. This makes data integration and governance a strategic, not just technical, consideration, echoing the approach Salesforce has already taken with Informatica's headless data management platform. Increasingly, the return on AI investment depends on the maturity of the data infrastructure underneath it, not just the sophistication of the agents themselves.
Looking ahead
The Salesforce-Databricks expansion fits a broader pattern, also visible in Salesforce's bet that enterprise systems should be built for agents rather than dashboards, where attention is moving from what AI models can do to what it actually takes to deploy them reliably inside large organisations. Trusted, governed and accessible data is emerging as a prerequisite, not an afterthought, for enterprise AI success. Salesforce and Databricks are betting that solving this problem will unlock the next wave of enterprise AI value, and customer experience teams are likely to be among the first to find out if that bet pays off.