Databricks acquires Tecton to boost real-time AI agent data

Databricks has announced that it has acquired Tecton, a company specialising in real-time data solutions for enterprise AI agents.

Purpose of the acquisition

This acquisition aims to provide enterprises with fast, reliable data access, supporting deployments of AI agents in demanding use cases, such as fraud detection, risk scoring, and personalised recommendation systems. By integrating Tecton's solutions, Databricks is focusing on enhancing its ability to deliver production-grade AI capabilities for enterprise customers.

Tecton has established itself as a prominent real-time enterprise feature store. The company centralises and automates the creation, sharing, and delivery of fresh, relevant contextual data for both classical machine learning and AI agent systems. Its platform allows for sub-10 millisecond latency, sub-100 millisecond freshness, and claims a 99.99% uptime, which is crucial for maintaining data accuracy and service availability at scale.

Tecton background

Tecton was founded by the creators of Uber's AI and machine learning platform, a foundational system that enabled Uber to support thousands of models in production. Tecton's technology provides a framework for enterprises to define, create, and share the data their models require, applicable across both historical and real-time use cases. This streamlines the often complex data preparation process and supports the reliable operation of AI agents in live production environments.

Describing Tecton's impact, the press release states: "Tecton makes it seamless to launch decision-making AI agents in production by preparing, curating and serving critical AI context needed to build customised and personalised agent systems. This marks a major step towards real-time production-grade AI for enterprise."

Data access for AI agents

One of the challenges in deploying AI agents in real-world enterprise scenarios is providing those agents with up-to-date, reliable access to a wide variety of data sources. This is particularly relevant for high-value use cases like fraud detection, where agents require immediate information about transactions, risk profiles, and user behaviour to make timely decisions.

"The main challenge is the ability to turn data from various data lakes, data warehouses, data APIs and streaming platforms into rich context available for AI agents in real-time. Without a clear system, preparing this data can be slow, repetitive, and prone to errors, making it tough for enterprises to move AI agents into production and slowing down innovation."

By automating the creation and delivery of this context, Tecton aims to minimise delays and potential errors that can result when enterprise teams manually assemble and update the data required for such applications. This helps reduce the risk and complexity associated with deploying production AI systems.

Integration benefits

The integration will bring Tecton's real-time data serving capabilities to Databricks' unified data and AI platform, including its Agent Bricks offering. This unified approach is expected to streamline workflows for customers, from raw data ingestion to deploying finished AI agents.

According to the release, "Bringing Tecton into Databricks will unite the best in online data serving with Databricks' Agent Bricks, empowering customers to build, deploy and scale AI agents faster and more confidently than ever before. Soon, customers can expect deeper integration, with Tecton's capabilities embedded directly into Databricks workflows and tooling, streamlining the journey from raw data to production AI agents."

The companies have an existing relationship, with shared customers in a range of industries. Tecton's systems are already used by both large corporations and startups with demanding AI applications. The acquisition formalises their collaboration and will provide Databricks' customer base with improved end-to-end capabilities for building and running AI agents.

The release also notes, "With Tecton, Databricks customers will gain fully integrated, automated online data serving within their unified data and AI platform. Most importantly, by uniting Tecton's industry-leading real-time data serving with Agent Bricks, customers will be empowered to build, deploy and scale AI applications faster than ever before."

Future expectations

Tecton's solutions, engineered to handle the full data lifecycle from definition to experimentation and production deployment, are described as reducing development complexity and providing a cost-efficient path to enterprise-scale AI adoption. Databricks and Tecton together plan to offer customers a streamlined approach for enabling both classic machine learning and agent-based AI applications, supported by reliable and timely data delivery.