Founder's Message

I have spent much of my career working close to analytics and data transformation.

Previously, I co-founded Diyotta, where I worked directly with enterprises modernizing their data platforms. This included moving to cloud data warehouses, rebuilding pipelines, and supporting teams that were trying to make analytics faster and more accessible. Diyotta was later acquired by ThoughtSpot, which gave me the opportunity to see analytics challenges at a much larger scale and across many organizations.

Through all of this, one issue kept showing up.

Database technology improved rapidly. Warehouses became more powerful and easier to manage. New tools emerged to help with transformations and reporting. At the same time, data modeling slowly became something teams tried to avoid or postpone.

I saw BI and analytics teams spend months rebuilding models. Joins, metrics, and business logic were often recreated inside BI tools rather than defined once and reused. Even with modern tooling, modeling remained manual, time-consuming, and difficult to maintain. Each new use case or new team added more work and more inconsistency.

I was closely involved in early validation and adoption efforts around dbt and Looker semantic layers. These tools helped bring discipline and structure to analytics workflows. However, modeling still required deep technical effort and remained tightly coupled to individual tools. Business logic had to be rebuilt and maintained repeatedly, and it struggled to keep pace with how quickly businesses changed.

What became clear to me was that analytics systems lacked business understanding. Data existed, but the meaning of that data lived in people's heads, documents, and dashboards. Machines could process data, but they did not understand how a business defined its concepts or metrics.

For a long time, automating this problem did not feel realistic. Traditional AI and machine learning could infer schemas or detect patterns, but they struggled with business meaning and context. Data modeling is not only technical. It is semantic.

The arrival of GenAI changed what was possible. GenAI made it feasible to reason across schemas, metadata, usage patterns, and business language together. This opened the door to automating parts of data modeling while still keeping humans involved for review and governance.

That is how Codd AI started.

Codd AI focuses on automating data modeling and semantic foundations using AI, while keeping accuracy and trust at the center. The platform generates ontologies, relationships, and metrics, and allows teams to review and certify them. This creates a shared foundation that BI tools and AI systems can rely on.

This is a difficult problem, and it is not solved by a single feature or tool. But after seeing the same issues repeat across many organizations, it became clear that without a strong semantic foundation, analytics and AI cannot scale.

Codd AI exists to build that foundation.