Structured Finance’s Data Nightmare and A New Blueprint for Trust

I understood the problem at BNY and Moody’s and I found the solution in AI + Blockchain

Two Vantage Points, One Persistent Challenge

At Moody’s, as a ratings analyst, I saw how much deal certainty depends on data clarity, yet loan tapes often arrived in varied formats, with gaps that forced teams into repetitive clean-up.

At BNY Mellon, as a modeler, I felt the downstream effects: when upstream data isn’t standardized, modeling, waterfalls, and reporting slow down while people reconcile fields by hand.

The lesson that stuck with me was simple: when data is messy upstream, every downstream process pays the price.

The monthly clock everyone runs by

In many deals, performance data is reported between the 20th and 25th of the subsequent month. If scheduled payments weren’t tied to that window, much of this work would likely slip because the final reports often come together right up against the deadline. The cadence isn’t a lack of effort; it’s a reflection of the friction created when each participant is reconciling, versioning, and emailing their own view of the loan tape.

What We Still Hear in 2025

At a recent symposium of all key stakeholders in Residential Mortgage Backed Securities (RMBS), everyone agreed with the need to make data cleaner, faster, and more consistent.

But who will bell the cat? Trustees wanted to know, ‘even if were able to provide, who will bear the cost?’ Issuers had a similar response.

This is even before we got into the need for ESMA reporting requirements for Europe and Bloomberg reporting for public deals.

Many institutions have already invested heavily in traditional ETL (Extract, Transform, Load) tools plus consultants, and with every small change leading to incremental costs. Most technology teams cannot think beyond traditional ETL, so much so that on a recent client call, a participant kept asking us, “So what’s your ETL tool?”.

Likewise, industry groups like the Structured Finance Association (SFA) continue working on standards that can reduce variability at the source but the process focusing on one narrow subsegment of one asset-class itself has taken a lot of time.

Time industry consensus would take and investment traditional ETL approaches need, means trusted loan data, available in real-time will remain a holy grail.

There’s a complementary path enabled by AI that doesn’t require waiting for perfect standards or multiplying reconciliation workflows: clean at the source, once, and let every stakeholder benefit in real time.

A Third Path: AI to Map Data at source + Shared Digital Rail

Instead of reconciling after the fact, at Intain, we use AI to standardize and validate loan data at ingestion, across servicers and asset classes, so everyone works from the same “golden dataset” as early as possible.

AI mapping at ingestion:

At the core of this solution is ability to take loan data from the servicers in any form and format, use of AI to map to a master template. This is a one-time exercise at the time of servicer and deal set-up. We pair that automation with analyst controls (review, exception handling, overrides). at the time of servicer and deal set-up. We call it collaborative intelligence: machines for the repeatable mapping and checks; humans for due diligence.

Normalize field definitions and formats across tapes without any software upgrades required by servicers or administrators (trustees, facility agents)

Configurable Master Templates:

We derived the best practices from the ESMA templates, in terms of definition of fields and data checks. Then we extended it to cover all the nuances, and made the master template configurable.

At the time of the deal set-up, as you make the choices of asset class (e.g. Residential Real Estate), sub-asset class (e.g. Fix & Flips or HELOCs), and other attributes (e.g. [Prime, Sub-prime], [QM, nonQM]), a master template to map the tape becomes available.

Golden dataset in real time

Once the mapping is done, Intain can ingest data in real-time ensure standard loan tapes — a single version of the truth that issuers, trustees, investors, and rating agencies can rely on. This also enables real-time analytics, but that is a subject for another blog.

*Raw Tape

*Standardized Tape

Result is that for a fix and flip securitization, irrespective of the deal, issuer or servicer, there will always be a standard data format. Same is true for a HELOCs securitization, or an auto loans credit facility, or whole loan sale of RTLs.

Shared Through a Digital Ledger

Because this happens up front on a shared digital rail, you don’t multiply the same reconciliation effort across counterparties; you remove it at the source.

Every step, raw tape as received, mappings applied, validations performed, and calculations used for reporting, is recorded on a digital ledger. That gives trustees, rating agencies, and investors a consistent audit trail without emailing spreadsheets around. It also means oversight isn’t bolted on after the fact; it’s built into the data’s lifecycle. It ensures auditability that travels with the data

What changes for each stakeholder

• Issuers gain speed to market with fewer back-and-forth cycles.
• Trustees reduce manual reconciliations and have an immutable record of actions and data states.
• Investors get clearer line-of-sight into underlying assets, earlier.
• Rating agencies work from a single source of truth, making audits fast and transparent.

In more than 80 transactions across asset classes that Intain has onboarded, this approach has cut manual effort by up to 90% and reduced turnaround times by roughly half, while giving every participant a single version of the truth. Your mileage will vary by asset class and process design, but the direction is consistent.

In Summary

Here’s how we have used the latest in technology to address a challenge that the industry had thrown up its hand about:

1. We Handle Complexity: Our AI-powered system intelligently interprets and normalizes existing data formats on the fly. It understands the nuances between asset classes without needing hundreds of predefined fields.
2. We Eliminate the Cost Barrier: Our solution is offered as a service at a negligible cost, making it immediately accessible and removing the “who pays?” debate.
3. We Work with Existing Data: We don’t require servicers to change their systems or capture new information. Our technology works with the data they already have, structuring it for all stakeholders.
4. We Ensure Ongoing Consistency: Our process is automated and seamless for both closing and monthly servicing tapes, guaranteeing consistent, reliable data throughout the entire lifecycle of a transaction.
5. Powers Ida, first virtual analyst for structured finance: Addressing the data inconsistencies, allowed us to launch Ida, at the end of 2023.