Ten years ago, Databricks set out to unify analytics, data engineering and machine learning on a single platform. Fast forward to the Data + AI Summit 2025, and that vision has gone mainstream. 22,000 attendees in person. 65,000 more online. And product launches that are already transforming enterprise AI.
As someone leading data transformation for Fortune 500s across retail, automotive, CPG, travel and financial services at Valtech, I came to #DAIS25 to ask one specific question: Can these tools activate meaningful change today?
The answer? Unequivocally yes. Here’s what you need to know.
The lakehouse just became your database
At the summit, Databricks unveiled LakeBase, a groundbreaking transactional database that separates storage and compute while maintaining open formats.
This signals a major shift: transactional AI at enterprise scale. By decoupling storage and compute while staying open format and Postgres-native, LakeBase powers use cases that weren’t previously viable.
LakeBase is built on open source, runs serverlessly and supports instant database branching. That unlocks real-time applications with millisecond latency and high scalability.
For Valtech clients, the impact is immediate:
- Retailers and CPG brands can now deliver live product recommendations using unified behavioral and product data.
- Travel and hospitality companies can react to customer behavior in real time.
- Financial services companies can deploy real-time fraud detection with governed access.
From agents to agentic systems
Databricks is moving quickly toward production-grade AI agents. With AgentBricks, customers can now:
- Automatically evaluate LLMs using domain-specific judges.
- Optimize agent behaviors via 100-plus ML techniques.
- Make transparent tradeoffs between speed, cost and output quality.
These agents support continual learning, enabling systems that get smarter over time from user feedback.
Valtech is already building such agentic workflows. Imagine a customer service bot that updates BI dashboards, books follow-up actions and summarizes campaign outcomes. Now, we can do it faster, cheaper and more reliably with AgentBricks.
Apps, not just models
Databricks is transforming how enterprises interact with AI by launching its own Apps Platform, enabling developers to build custom, governed AI tools using:
- Streamlit, Gradio, Dash (Python)
- Node.js (JavaScript)
- One-click integration with LakeBase for operational storage
More than 20,000 AI apps have already been built on the Databricks Apps Platform since November 2024.
For Valtech clients, this opens a new frontier of embedded intelligence inside day-to-day operations. Imagine an automotive sales tool that surfaces real-time offers based on region, incentives and customer history. Or imagine a campaign assistant for CPG brands that adjusts copy based on past ROI.
This is the kind of integration that makes AI part of the business.
Unified governance that’s actually unified
With Unity Catalog, governance now covers everything: structured and unstructured data, models, dashboards, and metrics. It tracks lineage, applies business semantics, and supports both Delta Lake and Apache Iceberg.
For most enterprises, governance has been a blocker. At Valtech, we’re already helping clients move from fragmented policies to unified control. Unity Catalog makes that journey possible and scalable.
Data for billions, not just the few
Databricks is democratizing access to data and AI by launching a free edition of its platform and investing $100 million into AI education worldwide. All self-paced learning materials are now open source.
This democratization aligns perfectly with Valtech’s vision of empowering client teams and building lasting capability.
Real-world use cases
The future of AI is unfolding all around us. At DAIS, we saw how companies are moving from prototype to production-ready frameworks thanks to tools like DSPy on Databricks.
