Insights from Databricks at CData Foundations 2025

CData Foundations 2025 gave us a look into the evolving priorities of the Databricks ecosystem. With two insightful sessions led by Elise Georis (Senior Staff Product Manager at Databricks) and Eric Tome (Solutions Architect, Databricks) alongside Andrew Chabot (Senior Manager, Data Engineering at FinThrive – our mutual customer), the message was clear: the future of analytics is real-time, federated, and agentic.

Three recurring themes emerged from the sessions:

1. The importance of real-time data integration

2. Empowering domain teams with self-service pipelines, reducing time-to-value

3. Bridging the operational-to-analytic divide with federation and live connectivity

Let’s explore each of these themes in depth, to understand how Databricks and CData partner to power our mutual customers’ data foundations.

Real-time and agentic intelligence

In her session, Elise looked ahead to the future of AI and agentic systems, where software agents act on behalf of users in real-time. These agents require low-latency, programmatic access to enterprise data across systems. Her team is prioritizing ways to expose real-time access in serverless environments like Databricks' serverless Postgres and unify these experiences with Delta tables and Unity Catalog.

In their session, Eric and Andrew shared this emphasis on real-time data connectivity and stated that many organizations are beginning to build intelligent systems that react to current conditions (pricing, fraud detection, user engagement, etc.), where traditional batch pipelines fall short. High latency and delayed ingestion can become real operational risks in these scenarios.

Supporting this kind of intelligence demands a data architecture that enables connectors and pipelines to feed systems in real time and expose them programmatically to agent frameworks. The CData platform plays an important role here: with support for live data access, change data capture, and real-time APIs for agent-triggered queries, CData connectors provide an accessible, low-latency path to enterprise data stored in external SaaS and operational systems.

Combined with Databricks' support for AI-native workloads and data governance, CData connectivity enables businesses to build event-driven pipelines and inference flows with current, trustworthy data, allowing agents to take actions based on the latest state of the enterprise.

Self-service pipelines: from bottleneck to scale

Data democratization was a central theme for both Elise and Eric. Elise emphasized the importance of enabling line-of-business users to configure data pipelines without deep technical intervention and maintenance. The goal here is to abstract complexity, including schema drift, API changes, authentication, while providing observability and governance through Unity Catalog. She laid out the complexities of database ETL and made the business case for empowering organizational stakeholders to do this self-service:

Eric also underscored that the most successful data organizations are shifting to low-code or declarative tools that expose ingestion as a reusable capability. Teams should be able to select data sources and define pipeline logic without building or managing custom connectors every time. This abstraction of complexity enables organizations to reduce time-to-value and extract deeper, contextualized, and more actionable insights from data assets.

CData’s approach aligns well with this direction. By packaging complexity into prebuilt, no-code connectors and exposing external systems through intuitive UI and metadata models, CData supports fast onboarding, empowers self-service, and scales ingestion with minimal friction. This frees engineering teams from pipeline maintenance and enables domain teams to take ownership of their data needs.

Bridging operational and analytic systems with federation

Elise highlighted Databricks’ investment in Lakehouse federation and its ability to query data across systems without physically ingesting it. Federation lowers the cost of experimentation, accelerates time to insight, and prevents unnecessary data duplication. It also expands access to operational systems and cloud applications that may not be suitable for batch ETL (due to latency, cost, or compliance).

Along similar lines, Andrew shared how FinThrive's teams faced challenges integrating data from external systems, citing brittle ingestion pipelines and schema drift as persistent pain points that were addressed by CData automations. For many data sources, FinThrive worked with CData to establish real-time access via virtualization instead of moving data altogether. Eric stressed that pushdown query capabilities are crucial here - federation must still perform at scale.

CData helps address these concerns by supporting live connectivity with built-in federation and pushdown across systems, reducing the friction of moving and joining data across systems. This improves pipeline resilience, supports faster iteration, and enables real-time decision-making without the need for redundant storage.

CData + Databricks: accelerating time to insight

At Foundations 2025, Databricks speakers emphasized a clear direction towards enabling deeper insight and automation, spanning more operational systems across an organization’s data estate. CData helps Databricks customers achieve this through a managed, real-time connectivity layer that eliminates the bottlenecks of manual integration and custom connectors. Whether accessing external SaaS platforms, databases, or APIs, CData ensures that Databricks users can work with current, trustworthy data wherever it resides. This accelerates time to insight, reduces development and maintenance overhead, and enables more agile analytics and AI use cases.

Importantly, CData’s integration with Databricks is designed with platform consistency in mind. We align with Databricks-native services like Unity Catalog and Delta Live Tables, supporting a unified approach to data governance, lineage, and observability. Our connectors operate within the semantics of the Lakehouse, enabling live queries, real-time ingestion, and agent-triggered access - all while supporting fine-grained security and monitoring.

As enterprises build toward more complex workloads, they need a data platform that removes the friction associated with high-quality data access and preparation. Together, CData and Databricks establish an enterprise data access, storage and analytics foundation to drive results with data.

Ready to see CData in action?

Start building faster, federated, AI-ready pipelines today.

• Explore CData + Databricks integration

• Try our 30-day evaluation of CData Sync