AtScale named OpenAI Select Partner to boost AI accuracy

AtScale has been named an OpenAI Select Partner, placing the data and analytics software company in OpenAI's partner network.

The partnership is intended to bring governed metric computation and shared business definitions into enterprise deployments of OpenAI models. AtScale argues this is needed to improve the consistency of answers produced by AI systems used for business analysis and decision-making.

The move extends AtScale's work linking OpenAI models and Codex with what it describes as a universal semantic layer. That layer is designed to give AI systems access to common definitions for measures such as revenue, margin and churn, rather than allowing separate tools or agents to calculate them differently.

AtScale recently published research that found leading AI models answered fewer than four in 10 business questions correctly without a semantic layer. The finding supports its argument that business context and common data definitions remain a barrier to wider corporate use of generative AI tools.

Business context

AtScale's products are used by enterprise AI and data teams for semantic modelling, metric computation, query optimisation and governed data access. Its work with OpenAI includes connecting Codex to its semantic layer so coding agents can use established business definitions in applications, dashboards and analytical workflows.

The company also highlighted two parts of its product set in the announcement. AtScale Ava is designed to help clarify ambiguous business questions, while its AI Computation Engine calculates metrics using shared definitions and calculation logic.

Chris Lynch, chief executive officer of AtScale, said the company sees a shift in how businesses are approaching AI projects.

"Enterprise AI is moving from experimentation into everyday business decisions, and enterprises cannot have every agent inventing its own version of revenue, margin, or churn," Lynch said. "Being named an OpenAI Select Partner reflects the growing demand for trusted business context and gives us an opportunity to help more organizations close the enterprise AI accuracy gap."

Agent accuracy

The partnership also reflects a wider issue for companies seeking to use AI tools beyond narrow pilot projects. As more organisations deploy models across finance, operations and customer functions, it has become more important that those systems use the same definitions as internal reporting.

Jay Schuren, president of AtScale, said the company's work is focused on preventing differences in business logic between departments, dashboards and AI tools.

"An agent's only as good as the business context it's given," Schuren said. "Our work with OpenAI connects its frontier models and Codex to AtScale's shared business definitions, so every metric is computed the same way, every time. When a coding agent builds a dashboard, it should use the same definition of revenue the CFO signs off on. That's how we help customers build agents they can trust with real business decisions."

OpenAI's partner programme brings together companies that build, sell and deliver AI products and services based on its models and tools. AtScale's inclusion as a Select Partner signals a closer commercial and technical relationship as both companies target enterprise customers seeking more reliable use of generative AI in analytics and workflow tasks.

AtScale's role centres on connecting AI systems to governed business logic rather than leaving models to infer meaning from raw data alone. It says this approach is intended to help organisations measure results, identify gaps in business context and improve the reliability of outputs used in operational and analytical settings.