Veritas Data Research goes AI-native with Claude
Source Meridian and Veritas Data Research are working together to move Veritas to an AI-native way of working: a Claude-based development workflow for its engineers, and a shared context lake that lets every department reach the knowledge it needs through Claude.
Who they are
Veritas Data Research, based in the United States, builds high-fidelity vital status and provider information datasets that close critical gaps in healthcare data. Its products include the Veritas Fact of Death Mortality Index, Veritas Cause of Death, and the Veritas Provider Network and Affiliations Index, which covers more than one million U.S. physicians. Veritas serves healthcare providers, life sciences companies, health plans, public health agencies, researchers, and financial services and life insurance firms.
Those same industries are increasingly building AI models on health data, and those models are only as accurate as the mortality signal behind them. A model trained or validated against a stale or incomplete death record treats someone as alive after they've died, and every prediction downstream inherits that error. Veritas gives AI teams, and researchers validating their outputs, a current, audited record of who has died.
The Challenge
Veritas built software the traditional way: engineers wrote code by hand through a conventional development lifecycle. The knowledge that keeps a data company running, such as how each team works, its processes and its definitions, sat inside each department, where other teams couldn't easily find or reuse it.
Veritas wants to become AI-native: meaning faster delivery for its engineering team, and a way for every department to work from the same understanding of the company
What Source Meridian Built
Source Meridian worked with Veritas's engineering team to build an AI-native development workflow around Claude Code. That work is the leading edge of a broader shift already underway at Veritas, where the same approach to Claude now reaches every department through Claude Cowork and Claude Design.
-
An AI-native development lifecycle.
Veritas's engineering team is moving from hand-written code to a development workflow built on Claude Code, and is now delivering software up to 5x faster.
-
The company-wide harness behind it.
That engineering work is part of a bigger effort: Veritas is building a single harness around Claude, company-wide. Juan Felipe Hernandez, Tech Lead and Architect at Source Meridian, is leading that work. It gives every Veritas department the same company context, plus skills built for that team's work: Customer Support, Data Science, Design, Engineering, Global, Marketing, Product, Sales, Senior Leadership and Admin. Every team works with Claude from a common understanding of what Veritas is and how it works.
-
A context lake across departments.
Each department's knowledge about its own processes is captured in a shared context lake, accessed primarily through a Veritas marketplace plugin. MCP connectors extend this by pulling in context from Veritas's wider tools, such as Confluence and SharePoint. People in one department, and Claude working on their behalf, can pull what they need from another department without waiting on a handoff.
Outcomes
5x faster software delivery
10 departments working from one shared harness
5 skills, 1 plugin & 4 MCPs in production
15 people using Claude every week across Veritas
See what an enterprise context lake can do for your organization.
Let’s discuss how customized Claude workflows can accelerate your software engineering while preserving complete data integrity.