Databricks & OpenAI: Enterprise AI Gets Real

At the Data + AI Summit 2026, the partnership between Databricks and OpenAI took center stage, signaling a significant push to make complex AI, particularly agentic AI at scale, a production reality for enterprises. The core message was clear: the era of AI prototypes is giving way to operationalized intelligence, with data context and governance now paramount.

The collaboration positions Databricks as the enterprise data and AI fabric, providing the necessary context, governance, and infrastructure. OpenAI, in turn, supplies its cutting-edge intelligence and foundational models. This synergy aims to transform nascent AI experiments into reliable, production-ready agents. According to Databricks, the event underscored a shift where intelligence is no longer the bottleneck; data ground truth and context are the new priorities.

From Research to Production

The partnership is accelerating efforts to move AI from research labs to enterprise workflows. A key highlight was the native integration of OpenAI's models, including GPT and Codex, within the Databricks platform. This allows organizations to build custom AI agents with built-in governance, controlling access, actions, and costs.

Agent Bricks serves as the developer platform where OpenAI's intelligence meets Databricks' enterprise control. This platform aims to streamline the development of AI agents by addressing the often-overlooked 99% of work involved in deployment, security, monitoring, and context management, beyond just the core model.

Governing AI at Scale

A critical component of making AI production-ready is robust governance. The Unity AI Gateway is designed to ensure trustworthy AI deployments at scale. It provides auditing for every model interaction, enforces per-user budgets, and offers security and compliance teams the visibility needed for broader AI adoption.

Databricks Agent Tools, integrated into Agent Bricks, grant OpenAI-powered agents secure, governed access to enterprise data. This ensures agents operate with business-relevant context, leading to more accurate decision-making. Leveraging OpenAI on Databricks promises faster time-to-value, reduced costs, and improved agent quality compared to integrating disparate tools manually.

Real-World Impact and Future Vision

The summit showcased tangible use cases. OpenAI co-founder Greg Brockman emphasized the opportune moment for builders, highlighting the rapid rise and native availability of Codex on Databricks. The Unity AI Gateway ensures that coding activities are auditable, cost-controlled, and compliant before reaching production.

Hertz Global presented two production use cases demonstrating the joint stack's impact. One involved an application built in 11 days using GPT-5.5 and Databricks to improve lead conversion rates from 60-65% to 75-80%. Another leveraged GPT-5.5 to categorize millions of customer feedback points in real-time, routing actionable insights to the appropriate leaders, significantly decreasing time to insight.

The conversation at DAIS 2026, including sessions with OpenAI CFO Sarah Friar and Databricks CRO Ron Gabrisko, reinforced that the true value of AI lies in redesigning work processes and embedding intelligence into core products for tangible business impact, not just usage metrics.

Databricks and OpenAI are hosting a joint virtual event, "Agents at Work: Shipping Agentic Apps at Scale," from August 4-6 to further discuss scaling AI agents in production.