Inveniam to Anchor DGML Extracted Data Elements On-Chain to provide verification for agentic workflows.

Docugami, the document intelligence company led by Jean Paoli, XML co-creator and former Microsoft open source subsidiary president, today announced it will open its Document Graph Markup Language (DGML), the technical foundation underlying its document intelligence platform. In a partnership announced alongside the move, Inveniam will anchor DGML-extracted data elements on NVNM Chain to make that data verifiable for AI systems. 

The release is aimed squarely at the agentic era. As AI agents take on more enterprise work, they increasingly act on data pulled from documents, but that data has been impossible to verify at a granular level. DGML converts complex documents into precisely labeled, query-able data points; anchoring those points on-chain lets an agent (or the human reviewing it) confirm provenance for an individual data artifact, not just a whole document. 

DGML extracts structure from documents without requiring training data or templates, a contrast with extraction approaches that depend on per-document-type model training. By opening DGML in July and supporting a developer community around it, Docugami aims to establish it as a shared, open foundation that asset managers, valuation firms, auditors, investors, and technology partners can collaborate with and build on.  

“The world's most important business decisions are made on the basis of documents that machines have never been able to read properly,” said Jean Paoli, CEO of Docugami. “We have spent years building the technology to turn complex documents into data with unsurpassed precision, and the next step is to make it openly available. By opening DGML, we are inviting every participant in the ecosystem to collaborate with us and build on a shared foundation.” 

“DGML is a foundational advance in how we read and structure documents,” said Patrick O'Meara, Chairman and CEO of Inveniam. “Anchoring DGML-extracted data elements on-chain ensures the data, once surfaced, can be trusted by every stakeholder, and every agent, that needs to use it.” 

NVNM Chain launched last month as Inveniam's purpose-built Layer 2, designed to serve as the receipts layer for agentic AI. It hashes and anchors verifiable data, including attestations such as Proof of Origin, Proof of State, and Proof of Process, so that AI agents and the institutions relying on them have a single, neutral record of exactly which data drove any decision or transaction. Anchoring DGML-extracted data on NVNM Chain places that document data directly into the substrate agents read against, making it both verifiable and audit-ready.