Ubie is a healthtech company on a mission to help people make more informed healthcare decisions with an AI-powered symptom checker for consumers and healthcare providers.
As a fast-growing healthtech company in Japan, data is at the heart of everything: from enabling smarter decisions across product, marketing, and operations to supporting the broader healthcare mission. But as the team scaled, so did the challenges of managing that data in a way that was governed, consistent, and self-serve.
"As our company grew, our data governance couldn't keep up. We needed a BI layer that gave us visibility, consistency and trust."
- Yu Ishikawa, Principal Data Architect
Ubie faced a common but complex problem: how do you give more people access to data, without losing control?
When only a handful of people use BI tools, it's possible to keep things working with best-effort practices and shared context. But once dozens of teams and hundreds of dashboards are in play, small cracks turn into systemic problems.
For Ubie, whose dbt codebase spans thousands of models and serves teams across regions and product lines, these issues became daily obstacles. Metrics definitions diverged between teams, business logic got duplicated across tools, and dashboards broke when data models changed. Non-technical users couldn't self-serve, and managing sensitive data access safely became increasingly difficult.
They needed a system that enforced consistency, caught errors early, and scaled access safely.
To support a fast-growing company with high data literacy (and a very lean data team), Ubie needed a governance model that could:
That's when they found Lightdash.
Ubie explored multiple BI tools including Metabase, Looker, and Preset. But Lightdash stood out immediately.
Unlike traditional BI tools that sit on top of your warehouse, Lightdash integrates directly with your dbt project. That makes it fundamentally different in how governance is applied: business logic is defined once in dbt, then reused everywhere without needing to reimplement it in the BI layer.
The dbt integration advantage
When you use dbt to structure your data warehouse, you're building what's called a medallion architecture. This approach organizes your data into layers starting with raw data, then cleaning and filtering it, and finally creating business-ready datasets that everyone can trust.
By implementing this medallion architecture with dbt, Ubie built reliable, tested business data layers that serve as the foundation for their analytics. Lightdash's dimensions and metrics layer sits perfectly on top of this foundation, allowing them to create a comprehensive metrics layer that's both maintainable and trustworthy.
This approach gives Ubie a powerful combination: dbt's strong data transformation capabilities working together with Lightdash's clear metrics definition framework. The result is a unified system where data transformations, business logic, and presentation layer all work in harmony, eliminating the inconsistencies and duplicated logic that plague traditional BI setups.
The AI integration advantage
Ubie saw something exciting in Lightdash's approach to organizing business data. When you define your metrics and dimensions clearly in Lightdash, AI can actually understand and work with your business logic in ways that weren't possible before.
Here's why this matters: most BI tools struggle with consistency. Different teams create different reports using slightly different calculations for the same metric. When AI tries to generate SQL from these inconsistent definitions, it gets confused and makes mistakes.
Lightdash fixes this by giving you one source of truth for all your business logic. Your metrics are defined once, stored centrally, and version-controlled. When AI needs to generate SQL, it's working from these reliable, consistent definitions rather than guessing.
For Ubie, this means they're building on solid foundations. As AI capabilities grow, they can adopt new features without sacrificing the governance and reliability that their healthcare operations demand.
Scalable governance capabilities
In addition to these advantages, Lightdash also stood out because of how deeply it supported scalable governance:
This combination of governed definitions, developer workflows, and programmatic control gave them a blueprint to scale safely.
Beyond the technical capabilities, the Lightdash team's commitment to rapid iteration and community collaboration has also been exceptional for Ubie.
"The quick release cycles and dedicated support demonstrate a team that truly understands enterprise needs. The team's responsiveness and willingness to implement community-requested features has created a collaborative environment that benefits all users."
- Yu Ishikawa, Principal Data Architect
Lightdash's project structure follows a clear, structured hierarchy that allows Ubie to easily govern access to data. Global rules are applied at the organization level, with further segmentation by product or function at the project level, and fine-grained control of dashboards at the space level.
Infrastructure-as-code governance
Ubie's approach to governance extends far beyond traditional BI administration. They developed a custom Terraform provider that integrates seamlessly with their existing approval workflows through GitHub pull requests. This infrastructure-as-code approach ensures that all data access changes are tracked, reviewed, and auditable, which is exactly what compliance requirements demand for their regulated healthcare environment.
The Terraform provider automates role assignments, space creation, and permissions management, meaning governance decisions are code-reviewed, version-controlled, and repeatable. This API-first approach enables systematic data access control that meets the strict regulatory requirements they face operating in both Japan and the United States.
Embedding governance into development workflows
Most importantly, Ubie embedded these governance tools into their development process. Every dbt PR gets tested automatically against Lightdash dashboards using lightdash validate. If something breaks, the pipeline fails long before it hits production. This continuous validation ensures that changes to the underlying data models don't silently break downstream dashboards—a critical safeguard for data reliability.
Governance used to be a limiting factor. With Lightdash, it became an accelerant.
And critically: governance didn't slow them down. It gave them the confidence to move faster.
"Before, business teams had to wait days for custom queries. Now they can build their own dashboards, knowing they're using the right logic."
- Yu Ishikawa, Principal Data Architect
Governance is often treated as a constraint - something you layer on after the fact. But at Ubie, it's the thing that unlocked true self-serve.
Ubie's approach shows that with the right tools and mindset, governance can be embedded into your workflow from the start. It doesn't require endless dashboards, duplicated logic, or gatekeeping.
It just requires the right foundation - and for Ubie, that's Lightdash.
"We don't just use Lightdash as a BI tool. We treat it as a governed layer on top of our dbt project that's ready for the AI-powered future. That comprehensive approach, from governance to community collaboration, is what makes it scale."
- Yu Ishikawa, Principal Data Architect