Eastman unifies data and builds an AI-powered future with Microsoft Fabric

Eastman touches nearly every aspect of modern life. From the food we eat to the clothes we wear to the medical devices we rely on when sick, Eastman’s materials and innovations help its industry customers tackle the world’s toughest challenges.

Delivering this kind of impact requires strong partnerships with customers, which is where the sales and support teams at Eastman shine. They work hard to build deep relationships, and a big part of this is creating detailed reports after every interaction. 

Eastman had a lot of customer data across various domains, but the company’s monolithic architecture created dependencies that held back data development teams. This also made it difficult to extract additional insights for search and integrate them into modern applications. So, when sellers prepared for meetings, they spent hours sifting through documents, spreadsheets, and call summaries. Important information, like pricing discussions or contract renewal dates, could be buried under several months’ worth of notes. “Instead of having a single hub for this preparation, we had thousands of data points to look through—which created a lot of friction with our processes,” says Andrew Ervin, Manager of Generative AI at Eastman.

Building a modern data architecture

Eastman wanted sellers to find information quickly, but the company’s legacy data warehouse—built on a traditional ETL model with SQL Server—offered limited scalability and wasn’t designed for large volumes of semi-structured or rapidly changing data. “We have a massive amount of data, and it’s always changing,” says David Purdy, Senior Associate Enterprise Architect at Eastman. “That’s why we moved to a cloud platform with built-in analytics.”

To unify data, Eastman adopted Microsoft Fabric and moved toward a distributed, domain-oriented architecture. Teams now ingest and manage data within their own workspaces while using OneLake shortcuts to make datasets available across domains without unnecessary duplication. In practice, Eastman replicates critical source datasets into Fabric (data mirroring), transforms the data using dbt, and loads it into a warehouse to be exposed to other teams via shortcuts, preserving governance and reducing copy sprawl. “If I need data from, say, finance or supply chain, I can mirror that data and shortcut that into my workspace to build applications on top of it,” Purdy explains.

For laboratory data, Eastman leverages open mirroring to ingest data from its lab information management systems into Fabric, uses dbt to transform and load it to a Fabric warehouse, and then exposes the curated models to analysts and data scientists. The company has ingested roughly a billion rows from eight systems so far and runs hourly updates, enabling both operational reporting and large‑scale machine learning that weren’t feasible on the old platform.

Weaving AI into the company fabric

With its data unified on Fabric, Eastman finally had the analytics and scale to reimagine sales prep and customer research. The company built a sales copilot directly into its CRM using Azure Web Apps, Azure OpenAI, and SQL database in Fabric. Essentially a personal AI assistant, the copilot can retrieve information or answer specific questions in seconds. “We wanted to give our sellers actionable answers—that will meaningfully drive customer decisions—in as few clicks or keystrokes as possible. So, we made the app voice-first,” explains Ervin. Now when sellers prepare for meetings, they just speak into their device. The copilot converts the voice input into a SQL query that filters the database and extracts the most relevant information. The results appear in a dynamic dashboard with text, graphs, and visuals. Sellers can keep refining the dashboard by talking to the copilot, saying things like ‘Two meetings ago, they talked about sustainability—put that in my dashboard.’  

According to Logan Finke, Principal AI Data Architect at Eastman, having SQL querying capabilities is critical to surfacing such hyper-precise information: “If you’re asking something like ‘how many inquiries have we had in the last few months?’, that’s a SQL query—not a vector search. But we need vector to identify where inquiries happened in the first place. We chose Fabric because it does both.” Fabric is used to process the data and generate vector embeddings, which are stored using SQL database in Fabric. GraphQL then acts as an integration layer to help the copilot surface insights. For GenAI developers, GraphQL simplifies data integration by allowing applications to request the exact data they need from a database or backend. “With Fabric, we can handle extraction, transformation, loading, and generate insights all in one place—that’s a tremendous win for us,” says Finke. 

From four hours to forty minutes

The sales copilot is also slashing time spent on reporting. Sellers just talk to the agent, which writes the report and pulls out key customer details. Azure OpenAI then cleans and organizes the data before sending it back to the CRM. Finally, it’s fed back into Fabric, ready for the next person to analyze. “Our sellers no longer need to chase down data from other departments. They just ask one or two questions to get what they need,” says Ervin. And because Fabric integrates with Microsoft Entra authentication, Eastman can set role-level policies to ensure only approved employees have access to trade secrets or sensitive research. 

Eastman is still early in its rollout of the sales copilot, but initial results are impressive. Sellers who once spent up to four hours preparing for calls, can now do it in approximately 40 minutes. 

A future built on insights and innovation

Unifying its data on a platform like Fabric has been transformational for Eastman. Sellers are saving hours preparing for calls and writing summary reports—and they’re having better, more insightful conversations. That’s helping Eastman build stronger, more strategic relationships not just with individual customers, but entire market segments. 

“We operate in a lot of markets that are fundamentally different from one another. Being able to aggregate all this loose, unstructured data into something that’s actionable is really helping our commercial organization build better strategies for the year ahead,” says Ervin.

On the data warehousing side, Eastman can now scale up and out indefinitely—a capability Purdy calls “game changing.” The company sees the future of data eventually shifting away from dashboards and reporting to natural language interfaces. Eastman is already embracing this future with its own sales copilot, and plans to keep innovating. As Finke sums it up, “There’s a lot of innovation happening at Eastman, and we need a data and analytics platform that moves at our speed—and that’s Microsoft Fabric.”