Why I Built This: The Procurement Data Problem Nobody Wants to Audit

I spend four days a week auditing defense contractors for the Defense Contract Audit Agency. My job is to evaluate whether the costs a contractor bills to the federal government are allowable, allocable, and reasonable. That means I spend a lot of time inside procurement systems — vendor masters, purchase orders, subcontract files, invoice trails.

The patterns I see in defense contracting are not unique to defense contracting.

What the data actually looks like

A mid-sized defense contractor with $800M in annual revenue might have three or four ERP systems — one from a legacy business, one inherited through an acquisition, one standing up a new division. Each has its own vendor master. None of them talk to each other in any meaningful way.

The procurement team knows this. They work around it. They maintain spreadsheets that bridge the gaps. They have institutional knowledge about which system holds the authoritative record for which class of spend. They have people who know, from memory, that the vendor in System A called “Allied Industrial” is the same entity as “AIX Corp” in System B.

When one of those people leaves, the knowledge walks out with them.

This is not a defense contracting problem. I have seen the same pattern in every multi-entity business I have worked with or studied: industrial conglomerates, PE-backed portfolios, holding companies that have grown through acquisition. The architecture is the same. The workarounds are the same. The institutional dependency is the same.

The audit exposure

From an auditor’s perspective, the risk is straightforward. If you cannot produce a clean, traceable record of who you paid, for what, under which contract, at which price — you have an exposure. For a defense contractor, that exposure is a finding. For a public company, it is a controls deficiency. For a PE-backed portfolio, it is a valuation haircut when the next buyer’s due diligence team arrives.

The exposure is not fraud. It is almost never fraud. It is disorganization that looks like fraud when audited, because the records that would prove legitimate business purpose either do not exist or cannot be produced in a usable form.

Why existing tools don’t solve it

There are spend analytics platforms that promise to consolidate procurement data. Most of them work reasonably well when the underlying data is clean and the company’s ERP footprint is a single instance. They struggle when the ERP footprint is fragmented, because they were built on the assumption that the data layer is already resolved.

Entity resolution — the process of determining that five vendor master entries represent one legal counterparty — is not a dashboarding problem. It is a data engineering problem. The vendors that sell dashboards are not paid to solve data engineering problems. They are paid to make dashboards.

This creates a gap. The data engineering work that would make procurement analytics trustworthy is unbundled from the analytics tools that procurement leaders actually buy. The analytics tools get purchased. The data engineering work does not get done. The dashboards surface numbers that nobody fully trusts, so nobody fully acts on them.

What RDMIS is

RDMIS LLC is the answer to a question I kept asking in the field: who is doing the data engineering work that the analytics vendors assume has already been done?

The answer, consistently, was nobody — or someone’s overworked data team, doing it manually, inconsistently, without a repeatable process.

Lucint is the platform I built to do that work systematically. Forseti handles the internal layer: entity resolution, price variance detection across brands, contract renewal radar. Yeti handles the external layer: import manifest intelligence, vendor hiring signals, patent velocity tracking. Safe Harbor is the architecture that ensures raw contract data never crosses brand boundaries — because in a multi-brand environment, one brand’s pricing should not be visible to another brand’s procurement team.

This is not a new category. It is a gap that exists because the market organized itself around dashboards before it organized itself around data quality. RDMIS exists to close that gap.

I am still a federal auditor. The exit condition is a signed MSA with an anchor client. Until then, this is what I do with my Mondays.