Inveniam, Docugami Partnership Brings On-Chain Checks to RWA Data for AI
Private-market documents are getting a stricter data check. Inveniam and Docugami are working together to verify document data on-chain, mainly for AI systems that need better inputs than scanned PDFs, leases, loan agreements, and awkward contract language usually offer. Inveniam points to market analysis showing tokenized real-world assets growing from $14.1 billion to more than $32.4 billion this year. That is not a side note. That is a lot of money sitting on paperwork, and I’ll be honest: I would want the paperwork checked too.
Inveniam says it already anchors more than $200 billion in private-market assets on-chain. With Docugami, it plans to verify document data at the element level for real-world assets. The issue is blunt. AI can be useful, but it still trips over unstructured data buried in legal and financial documents. Most guides say the answer is better models. That’s only half right. The companies want to turn those documents into machine-readable data, then record the relevant pieces on a blockchain. Not only the file. The individual data points. The work will use Docugami’s Document Graph Markup Language, or DGML, and Inveniam’s NVNM Chain.
The deal also fits Inveniam’s plan to control more of its blockchain stack, according to company statements. Inveniam plans to acquire $MANTRA, the regulated Layer 1 blockchain that supports NVNM Chain as a Layer 2. The acquisition is expected to close by the end of the month. It follows Inveniam’s $20 million strategic investment in $MANTRA in August 2025. If the deal closes, Inveniam would own more of the infrastructure behind its RWA tokenization and verification work. Why does this matter? Because institutions usually care less about crypto slogans and more about whether the data underneath an asset can survive an audit. My take: this is an infrastructure story wearing a tokenization headline.
DGML is Docugami’s part of the system. It was co-created by Jean Paoli, a co-author of the XML 1.0 standard and former president of Microsoft Open Technologies. DGML turns documents such as leases, loan agreements, and valuation reports into labeled data elements. Inveniam’s NVNM Chain records those elements as time-stamped, tamper-evident artifacts on-chain. A document hash can prove that a file existed. This setup goes further by checking a specific rent number in a lease clause or a loan-to-value ratio in an underwriting memo. That is the useful part. Counter to the usual advice, the flashy AI layer is not the hard bit here. AI in finance has always had the same ugly problem: bad input produces bad output, only faster.
“The world’s most important business decisions are made on the basis of documents that machines have never been able to read properly,” Jean Paoli, co-creator of DGML, said. He said Docugami has spent years building technology to turn complex documents into data with “unsurpassed precision.” By opening DGML, the company wants more private-capital firms to use the format. Docugami says its technology uses open-source large language models, smaller reasoning models, knowledge graphs, and less template-heavy extraction to convert business documents into structured data without heavy template work or large training sets. Is that enough by itself? No. It may cut down on manual extraction errors, though I would still want people reviewing the important fields.
The RWA market has grown fast, from $14.1 billion to more than $32.4 billion, but data quality is still the weak spot. A tokenized asset is only as good as the data behind it. In the past, investors often had to trust the issuer or pay for separate due diligence. Inveniam launched NVNM Chain on May 7 as a Layer 2 for private markets and an attestation layer for agentic AI. It records dataset existence and issues Proof of Origin, Proof of State, and Proof of Process attestations, showing which data supported a decision or transaction. With DGML added, AI agents can read document data that can also be verified and audited. Patrick O’Meara, Inveniam’s chairman and CEO, called DGML a “foundational advance” for structuring private capital documents. Big phrase. The practical point is clearer: private-market data needs a better paper trail.
What this means
The RWA sector is moving beyond basic tokenization and into data infrastructure. That is less flashy, but probably more important. Tokenizing an asset does not help much if no one trusts the rent roll, loan memo, valuation report, or supporting contract data behind it. Yes, this sounds like it contradicts the tokenization hype. It should. Verifying specific fields inside private-market documents could make these assets easier for institutions to review, price, and audit. If the system works, it could support more capital moving into RWA protocols and Layer 1s such as $MANTRA.
Investors should watch two things: whether the product becomes publicly available, and whether Inveniam closes the $MANTRA acquisition. The DGML and NVNM Chain integration could become a model for document-level verification in tokenized assets. TVL in RWA protocols is worth tracking, but actual institutional usage matters just as much. A rising token price is noisy. A bank or asset manager using the system in production would say more. That’s the signal I would watch.
FAQ
Q: What is the main goal of the Inveniam and Docugami partnership?
A: The companies want to improve data trust for AI in the RWA market by verifying private-market document data on-chain and making complex documents machine-readable.
Q: How does Docugami’s DGML technology work?
A: DGML turns complex documents into labeled data elements using open-source language models and knowledge graphs. Inveniam’s NVNM Chain can then record those elements on-chain.
Q: Why does Inveniam’s planned $MANTRA acquisition matter?
A: The acquisition would give Inveniam more control over the blockchain infrastructure behind its RWA tokenization and verification work.
Q: How does this partnership address the “garbage in, garbage out” problem in AI?
A: It gives AI systems cleaner document data and makes individual data points easier to verify and audit.
Q: What could institutional investors gain from this?
A: They could get better data provenance and audit trails for tokenized assets, which may reduce some risk around private-market data.
Q: How could this affect the RWA tokenization market?
A: If institutions trust the underlying data more, they may be more willing to use RWA protocols and related Layer 1 networks.
Q: What is NVNM Chain’s role?
A: NVNM Chain records DGML-extracted data as time-stamped, tamper-evident on-chain artifacts and issues Proof of Origin, Proof of State, and Proof of Process attestations.
Q: Who co-created Docugami’s DGML?
A: Jean Paoli co-created DGML. He also co-authored the XML 1.0 standard and previously served as president of Microsoft Open Technologies.
Q: What documents can DGML process?
A: DGML can process business documents such as leases, loan agreements, and valuation reports, turning them into structured data.
Q: What is the current value of the tokenized RWA market?
A: Inveniam cites market data showing tokenized RWAs growing from $14.1 billion to more than $32.4 billion.