The Snowplow team recently had the privilege of joining over 15,000 data leaders, engineers, and AI practitioners at Databricks' Data + AI Summit in San Francisco. Of course, much of the conversation was focused on the explosion of GenAI tools and frameworks. But the bigger takeaway for us was this: 

The data stack is evolving from a system of record into a system of action—and it’s increasingly agentic.

By this, we mean data no longer just sits in reports for humans to read and act on later. Now, we’re entering an era where data automatically triggers smart actions in real time, like instantly offering discounts when someone abandons their cart or alerting support when a customer seems frustrated. 

To support this transition, Databricks launched a series of exciting new features at its Summit, including: 

• Databricks Lakebase: This is a modern database that combines the best of two worlds–it can handle your day-to-day app operations (like processing orders). Plus, it can handle your data analysis (like sales reports) all in one place. The benefit of this is that companies can build AI apps faster and cheaper because they no longer need to maintain multiple complex databases or move data between them. 
• Databricks Agent Bricks: An exciting new tool that builds and optimizes AI agents for your business by simply describing what you want them to do. With Agent Bricks, you now go from idea to production-grade AI agents in hours instead of weeks. 

We also saw the launch of new marketing intelligence tooling and enhancements to Unity Catalog, Databricks’ data governance system. 

But what was clear from the Summit was that the modern Lakehouse is being rearchitected for real-time workloads, AI-native use cases, and automated decision-making. 

Snowplow was proud to be part of several major product launches, customer showcases, and partner conversations throughout the week. Here’s a closer look at what we announced, what we learned, and how Snowplow is helping teams unlock real-time, AI-ready customer intelligence in the Databricks ecosystem.

Snowplow featured in Lakebase launch

Databricks introduced Lakebase, a new Postgres-compatible transactional database built directly on top of Delta Lake. It’s designed to bring operational (OLTP) workloads into the Lakehouse—eliminating the boundaries between transactional and analytical systems. As a result, Databricks customers can now build apps and run analytics on the same data without costly, complex integrations between different databases. 

The announcement follows Databricks’ acquisition of Neon, a serverless Postgres startup. This move mirrors a broader industry shift toward unifying operational and analytical workloads, similar to Snowflake’s acquisition of Crunchy Data.

At the Summit, Snowplow was delighted to be announced as a Lakebase launch partner. Snowplow allows organisations to stream rich, well-structured behavioral data into Delta Lake in real time—then immediately serve that data to operational systems. This eliminates the typical delays and data silos that prevent companies from acting on customer behavior instantly, enabling immediate personalization, fraud prevention, and smart recommendations. As a result, it’s now possible to drive higher conversion rates and reduce losses–all from a single, trusted data source. 

Supporting Databricks’ Data Intelligence for Marketing solution

We were also proud to be included in the launch of Databricks’ new Data Intelligence for Marketing solution. This offering brings together first-party data collection, governed modeling, and AI-powered activation into a single composable stack.

As part of the announcement, we showcased how Snowplow’s real-time event data enables marketing teams to:

• Build attribution models directly in Databricks
  
  
• Run customer journey and funnel analytics at scale
  
  
• Create dynamic Customer 360 based on behavioral conditions
  
  
• Power downstream personalization and engagement
  
  

By streaming high-fidelity event data into the Lakehouse, Snowplow helps teams move away from black-box CDPs and toward a more transparent, AI-ready foundation for marketing intelligence.

How Supercell modernised player analytics with Snowplow + Databricks

In his session, “Level Up Gaming Analytics,” Snowplow CEO Alex Dean shared how Supercell–the creators or games such as Clash of Clans– transformed its approach to player data to meet the demands of today’s live-service, cross-platform games.

Supercell needed to evolve from one-time reporting workflows to a real-time, experimentation-ready architecture. By adopting Snowplow’s behavioral data infrastructure and streaming events directly into Databricks, they’ve been able to:

• Maintain 100% data integrity, even during high-traffic game launches
  
  
• Unify anonymous and authenticated user journeys across platforms
  
  
• Run campaign experiments and player engagement analysis in real time
  
  
• Enable deeper community feedback loops across titles like Brawl Stars
  
  

Snowplow Signals + Agent Bricks: powering AI-native customer experiences

Databricks also introduced Agent Bricks, a framework for building, evaluating, and deploying intelligent agents that perform business tasks autonomously. These agents require high-quality context—something that Snowplow helps deliver.

In his talk, Snowplow Co-founder Yali Sassoon introduced Snowplow Signals, our real-time customer intelligence system built for customer facing AI agents and in-product personalization. 

Signals enables product and engineering teams to stream user attributes alongside historical user context from the Lakehouse—making it easier to drive relevant, dynamic customer experiences.

These real-time attributes can then be used to:

• Power next-best-action logic in applications
  
  
• Enrich ML models for churn prediction or LTV scoring
  
  
• Trigger personalization flows with full session awareness
  
  

As agents become more central to product and marketing workflows, the need for structured, interpretable customer context becomes essential. With Snowplow Signals and Databricks, teams can deliver on that requirement—without sacrificing control or governance.

Other Highlights From the Databricks Team

Alongside Lakebase and Agent Bricks, several other announcements stood out:

Unity Catalog Metrics
Databricks introduced support for certified KPIs directly in Unity Catalog. This allows teams to define key metrics—like session count, conversion rate, or time-on-site—as governed assets. For Snowplow users, this means modeled metrics from dbt can be reused across teams and tools with full lineage and governance. As a result, your teams can make faster, more reliable decisions using trusted data. 

Databricks One + AI/BI GenieDatabricks One and Genie were announced in preview as new ways to query data and build applications using natural language. These tools are designed to democratise insights across the organisation—but they depend on clean, well-structured data. Snowplow provides the semantic richness that makes these AI interfaces accurate, consistent, and valuable. 

MLflow 3.0 + Mosaic AI enhancementsFor teams building ML-powered applications, Databricks introduced updates to MLflow and Mosaic AI, including prompt versioning, agent observability, and storage-optimised vector search. These tools complement the streaming insights generated by Snowplow Signals—making it easier to train, evaluate, and iterate on models grounded in real user behaviour. So now, you can build better ML models faster using real, trustworthy user data. 

Looking ahead: building the next wave of intelligent infrastructure

One message came through loud and clear at DAIS 2025–the Lakehouse is evolving. 

It’s no longer just a platform for analytics —it’s becoming the foundation for real-time intelligence, operational decisioning, and AI-native execution. The modern data stack is moving from insight to action.

As systems become more context-aware and increasingly automated, the role of high-quality, behavioural data becomes even more strategic. At Snowplow, we see this shift unfolding across four key dimensions.

First, we’re enabling intelligent systems to make better decisions by delivering structured, contextual behavioural data. Our pipelines don’t just capture what happened—they preserve the how and why, with full session and entity-level awareness.

Second, we’re seeing more customers move beyond analytics and into operational, real-time decisioning. From powering product agents to triggering live interventions, Snowplow and Databricks together enable teams to embed intelligence directly into their applications—not just their dashboards.

Third, we believe that governance, transparency, and data quality are becoming foundational to responsible AI. Together with Databricks, we’re helping teams build infrastructure that’s not only performant, but trusted—through open schemas, data validation, and full control over the data lifecycle.

And finally, we’re continuing to support the builders—the data teams, engineers, and product managers designing systems tailored to their organizations. These are the teams choosing Snowplow and Databricks to move beyond one-size-fits-all SaaS tools and take ownership of their data strategy. We’re here to support them with the building blocks they need to scale, adapt, and innovate with confidence.

Want to learn more?
Explore Snowplow Signals or get in touch to learn how we’re helping Databricks customers unlock real-time, AI-ready behavioural data infrastructure.