Entrada

Data + AI Summit 2026 brought the Entrada team back to San Francisco alongside more than 31,000 members of the data and AI community. As a pure-play Databricks partner, being there felt like standing at the epicenter of the next chapter of enterprise AI.

A Few Proud Moments for the Entrada Team

The Data + AI Summit was a deeply meaningful week for us as a firm, and it was great to see clients and partners. We are also excited to share that we were named the 2026 Databricks Genie Partner of the Year. The award recognizes the work our teams have done to take Genie beyond the demo and into production — helping some of the world’s largest enterprises let business users ask questions of trusted, governed data in plain language and get accurate, analytical answers in seconds rather than days.

We were just as proud to see our own Kelly Zelenko named Databricks’ 2026 North America Technical Champion of the Year. Kelly’s technical leadership has been at the heart of some of our most complex and innovative Databricks AI initiatives.

An Agent-First Approach Takes the Stage

The 2026 Data + AI Summit was about putting agents to work. The headline for us was Genie One, a data-smart AI coworker grounded in enterprise deep context through the new Genie Ontology. Together they connect across an organization’s data to automate real data work using tools, skills, agents, and MCPs. 

Alongside it, Agent Bricks (now generally available and powered by the new open-source meta-harness Omnigent) gives teams an open way to build and run agents with shared context, agent memory, and secure, governed access to models and skills.

The ontology piece is the part we’d circle in red. Context is what separates an agent that genuinely helps from one that answers confidently and wrongly, and it’s precisely the foundation our Genie engagements are built on. Entrada is excited to help our clients stand up the trusted, well-modeled context layer that makes tools like Genie One reliable enough to act on.

Real-Time Meets Transactional: One Governed Lakehouse

Databricks closed a long-standing architectural gap this year. Lakehouse//RT brings millisecond speed and massive scale directly to the Lakehouse, removing the need to manage a separate system or copy data out for real-time analytics.

LTAP (Lake Transactional/Analytical Processing) goes further, giving users and agents a single governed system of record for both transactions and analytics in an open data lake. For mission-critical workloads, Lakebase added cross-region, cross-cloud disaster recovery, with one-step failover across workspaces.

The payoff for our clients: fewer moving parts and fewer copies of the truth to keep in sync. This is the kind of simplification that makes a governed, agent-ready estate far easier to design and maintain. 

Putting Engineering and ML on Autopilot

There were also a wave of announcements aimed squarely at productivity. Lakeflow unifies data engineering and lets Genie Code automate pipeline construction end to end, including through the visual Lakeflow Designer. Genie ZeroOps monitors production workloads, investigates issues, and proposes fixes for human verification, while Genie Code and ZeroOps for ML and Data Science extend that same automation to model training, optimization, and maintenance. 

For the builders, Genie App Builder brings “vibe coding” to the enterprise, turning natural-language intent into working data apps inside a governed environment.

The promise here is real: more output without sacrificing quality, with people verifying rather than babysitting. As a firm steeped in AI deployment, we see the biggest wins going to teams whose pipelines and models sit on a clean, well-governed foundation.

Governance That Keeps Pace with AI

Unity AI Gateway lets organizations control AI access, security, and spend for every person and every agent, leveraging per-user budgets, enforced guardrails, and automatic routing to lower-cost models. Unity Catalog itself has gone global, folding in the AI Gateway and spanning a customer’s entire Databricks footprint across accounts, regions, and clouds for one unified view of their data and its governance. 

And on the security front, Databricks continued building out Lakewatch, its open, agentic SIEM and security lakehouse, by announcing its intent to acquire Panther — an AI SOC platform whose 100+ pre-built integrations and detection-as-code will deepen automated threat detection and investigation once the deal closes.  

Agents Reach the Marketing and App Ecosystems

The agent story extended to the business teams, too. CustomerLake is an agentic CDP embedded directly in Databricks, giving marketers and data teams a workforce of agents to deliver the right customer experience at enormous scale. 

Apps on Databricks Marketplace now let organizations purchase, install, and run third-party apps directly inside their workspaces, within secure and governed environments. The platform is quietly becoming an app store. For our clients, these are powerful new ways to unlock value, but only when the data underneath is trustworthy.

New Era of Agentic Data and AI

Across every announcement, the same future shows up: one where agents are first-class consumers of enterprise data, reading your tables, following your relationships, and acting on the answers they produce. The platform keeps getting more powerful, but none of it rescues you from a badly modeled, poorly governed foundation; it only exposes it. Building that foundation is the work that earned us Genie Partner of the Year, and the work we love doing alongside our clients.

Wherever you are on your data and AI journey, everything Databricks unveiled this year only pays off on a trusted, well-modeled foundation. Want to know where your context and governance layer stands? Reach out for an ontology and Unity Catalog readiness assessment.