The Trust Layer for AI: Docugami and Inveniam Tackle Data Verification

KIRKLAND, Wash. and DETROIT – June 17, 2026 – In a move aimed squarely at one of the biggest unresolved challenges of the artificial intelligence boom, document intelligence firm Docugami announced it is open-sourcing its core technology. The company, led by XML co-creator Jean Paoli, is releasing its Document Graph Markup Language (DGML) and partnering with data orchestration firm Inveniam to create a verifiable, auditable data layer for AI agents.

The partnership addresses a looming crisis of confidence in enterprise AI. As autonomous agents are increasingly tasked with high-stakes work, they rely on data extracted from a sea of business documents—contracts, financial statements, and compliance reports. The problem is that this extracted data has been nearly impossible to verify at a granular level, leaving businesses exposed to significant risks from AI decisions based on flawed information.

Docugami’s DGML technology converts these complex documents into precisely labeled, queryable data points. Through the new partnership, Inveniam will anchor these individual data elements on its purpose-built NVNM Chain, a Layer 2 blockchain, allowing an AI agent or a human supervisor to cryptographically confirm the provenance and integrity of every single piece of data.

“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.”

The Anatomy of a Trust Problem

The transition to the “agentic era”—where AI systems act with increasing autonomy—is not a matter of if, but when. For CIOs, risk managers, and compliance officers, this presents a daunting challenge. An AI agent tasked with portfolio management could make a catastrophic trade based on a single misread number in a quarterly report. An automated compliance system might approve a transaction by misinterpreting a clause in a complex legal agreement. Without a mechanism to trust the underlying data, the promise of AI-driven efficiency remains hampered by unacceptable risk.

This is the data trust gap that Docugami and Inveniam aim to close. While many platforms can extract information from documents, their verification processes often rely on confidence scores or human-in-the-loop workflows. This new collaboration proposes a more fundamental solution: creating a tamper-evident, time-stamped record of each data point on an immutable ledger. This moves the verification process from a probabilistic assessment to a deterministic proof.

“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.”

A Two-Part Solution: Extraction and Attestation

The partnership’s strength lies in combining two distinct but complementary technologies. Docugami handles the sophisticated extraction, while Inveniam provides the cryptographic proof of integrity.

Docugami’s core claim is that its technology can extract structure from documents without requiring the extensive training data or rigid templates that define many competing Intelligent Document Processing (IDP) platforms. Instead of relying on layout-based models that falter when a document’s format changes, DGML uses a combination of open-source Large Language Models (LLMs) and small, proprietary reasoning models to build a semantic knowledge graph of the document. This approach allows it to understand the contextual relationships between terms, clauses, and tables, effectively learning a document’s inherent patterns in minutes, not weeks. This stands in contrast to established players like Google Document AI or Hyperscience, which often require significant configuration or pre-trained models for specific document types.

Once DGML has deconstructed a document into its constituent data atoms, Inveniam’s NVNM Chain steps in. Launched in May 2026, NVNM Chain functions as a specialized Layer 2 blockchain designed for attestation. For each data element—a specific date, a dollar amount, a counterparty name—a cryptographic hash, or digital fingerprint, is generated and recorded on the chain. This process provides what Inveniam calls “Proof of Origin,” “Proof of State,” and “Proof of Process.” This creates a permanent, auditable trail confirming where the data came from, what its value was at a specific point in time, and the process used to extract it, all without placing the sensitive data itself on the blockchain.

Echoes of XML: An Open-Source Playbook for an Industry Standard

The decision to open-source DGML is a strategic masterstroke, especially given Jean Paoli’s history. As a co-creator of XML (eXtensible Markup Language) at Microsoft, Paoli helped architect an open standard that became the bedrock for data interchange across the web and enterprise systems. By making the language of data description free and accessible, XML fostered a massive ecosystem of tools and developers, ensuring its ubiquity.

With DGML, Paoli is running a similar playbook for the AI era. By releasing the technical foundation in July and committing to supporting a developer community, Docugami is betting that a shared, open standard is the fastest path to widespread adoption. The goal is to create a common language that asset managers, valuation firms, auditors, and technology partners can all use to build trusted data workflows. This strategy sacrifices short-term proprietary control for the long-term network effects of becoming an industry-wide foundation.

This move has the potential to catalyze innovation far beyond Docugami’s own platform. If successful, an entire ecosystem of DGML-compatible tools for data validation, analysis, and agentic automation could emerge, accelerating the development of more reliable and compliant AI applications across the board.

Real-World Impact: Reshaping Private Markets and Beyond

The most immediate impact of this partnership will likely be felt in the opaque world of private market assets. Inveniam has already established a strong foothold here, anchoring data for over $200 billion in private-market assets. These markets, which include private equity, real estate, and infrastructure, are notoriously document-heavy and lack the data transparency of public markets.

This is particularly relevant for the burgeoning field of Real-World Asset (RWA) tokenization, which has seen its market value swell to over $32 billion. For a token representing a share of a commercial real estate building to be trusted and traded systematically, the underlying data—lease agreements, valuation reports, debt covenants—must be impeccably verified. The Docugami-Inveniam solution provides the foundational trust layer needed to make this a scalable reality, transforming illiquid assets into more accessible and machine-readable financial instruments.

While finance represents the initial beachhead, the implications are far broader. Any industry where critical decisions hinge on the interpretation of complex, unstructured documents—from legal and insurance to energy and aerospace—stands to benefit from a standardized method for producing verifiable data for their AI systems. This initiative is a critical step toward building an enterprise AI ecosystem where automated decisions are not just efficient, but fundamentally trustworthy.