Duodata published a new blog post on how Anthropic increased their agents' accuracy for answering analytics questions from 21% to 95% using a human-curated semantic layer.

Duodata gives AI a governed understanding of how your business works. The foundation is a Metrics Ontology: a connected model of the KPIs, drivers, and relationships that define business performance. It gives every KPI its full business context—what it means, why it matters, how it is calculated, what drives it, where its data comes from, how it can be analyzed, and who approves it. With that context, AI can do more than retrieve data or calculate a number. It can reason about business performance, explain why a KPI changed, answer questions using the correct business meaning, and recommend which levers could improve the outcome. Duodata governs this context centrally and keeps it aligned everywhere metrics are used across the company—from board reports, dashboards, and spreadsheets to semantic layers, data platforms, BI tools, and AI agents. Each system works from the same approved definitions and relationships, creating a consistent understanding of how the business is performing and what drives its results. The result is AI—and an organization—that understands the business and takes the right action to improve it.

Founded 2025