xMentium extracts data from complex documents to build the AI-ready data foundation enterprise intelligence depends on, discoverable in the OneLake Catalog and ready to use across Microsoft Fabric.
BARCELONA, Spain, Sept. 29, 2026 /PRNewswire-PRWeb/ -- xMentium, a Document Intelligence company that transforms unstructured enterprise documents into structured, AI-ready knowledge assets, today announced it is alongside IQ Sharing in Microsoft Fabric, a new data and intelligence sharing capability unveiled at the European Microsoft Fabric Community Conference (FabCon) in Barcelona.
The most valuable enterprise data rarely lives in a database. It sits in contracts, engineering drawings, P&ID diagrams, inspection reports, manuals, and correspondence, locked in formats that traditional pipelines cannot read. Extracting it has meant specialized connectors, custom pipelines, and thousands of hours of curation before a single insight reaches a business user.
IQ sharing in Fabric removes that barrier. For existing xMentium customers, it's the fastest path to the data they've already built. They discover xMentium in the Microsoft Fabric OneLake Catalog, connect and share using zero-copy technology the extracted output of their xTract pipeline, available as structured data, directly into their own OneLake tenant. Once in their own OneLake, customers can use the data with their favorite tools, like Microsoft Power BI, Microsoft Copilot, and Fabric Data Agents. The result is a durable, AI-ready data foundation rather than a one-off extract. Enterprises can build a curated, continuously refreshed layer of structured intelligence from their own documents and make it available across business units and partner ecosystems, all within Fabric and all governed by the policies they already trust.
"Across capital-intensive industries, the intelligence behind major investment and project decisions is buried in engineering documents, technical drawings, and decades of inspection and maintenance records, and that's exactly where generic AI breaks down," said Mike Kochanik, Vice President, Global Service Providers at xMentium. "RAG fails at the retrieval step when it can't identify the right document, version, or table. IQ sharing in Microsoft Fabric changes the economics of fixing that: customers discover xMentium in the OneLake Catalog and pull structured, governed data from their xTract extraction pipeline into their own Fabric workspace as zero-copy assets, so AI starts from a known fact instead of re-interpreting a scanned page on every query. The result is lower token cost, more deterministic answers, and value that accrues to the customer's operations, not the model."
"IQ sharing in Microsoft Fabric is about meeting customers where their data already lives and letting them bring more of it into OneLake without moving or copying it," said Dipti Borkar, Vice President, Microsoft IQ and OneLake. "xMentium helps organizations put their entire data estate to work for analytics and AI by turning unstructured documents into a structured, AI-ready data shared seamlessly via IQ Share that customers can discover and leverage in OneLake."
xMentium's integration with IQ sharing in Fabric will be available to Microsoft Fabric customers in the OneLake Catalog. Teams attending the European Microsoft Fabric + SQL Community Conference in Barcelona can see the capability demonstrated live. To learn more, visit xmentium.com.
About xMentium
xMentium helps organizations turn complex, unstructured documents into precise, structured information that people, systems, and AI can use.
xMentium is a launch partner for IQ Sharing in Microsoft Fabric, distributing AI-ready data to customers by offering a Document Intelligence processing engine, a foundational layer enterprise AI initiatives are missing. The platform handles the content that defeats generic OCR and off-the-shelf AI: scanned images, engineering drawings, P&ID diagrams, dense tables, contracts, and decades of operational and technical records. Every file gets classified and routed through a governed processing lane with available human review. What comes out is standardized output that is ready for enterprise systems and AI agents.
Media Contact
Rob Sheerr, xMentium, Inc., 1 (609) 500-5176, [email protected], https://www.xmentium.com
SOURCE xMentium, Inc.
Share this article