DataStruct AI is evidence-review software for compliance, quality and risk teams. It works through supplier files, audit records, contracts and registers against the requirements a team sets.
Where nothing supports a requirement, it says so. A named reviewer then confirms, rejects or flags each finding for follow-up, and that decision is recorded.
It is in private beta, sold directly, with limited places for founding customers and no self-serve sign-up.
Compliance, quality and risk teams ask the same kinds of question every cycle. Does this supplier hold the certificate we require? Was the calibration done? Did the board approve the statement? The answer is usually somewhere in a folder of contracts, certificates, registers and policies, written in words that are not the words of the question.
DataStruct AI's own account is that keyword search misses it, and that general-purpose chat tools answer confidently whether or not the document says so. It was built for the space between them, which is finding the evidence, showing where it came from and being clear when it is not there.
The core of the product is an Evidence Programme. A named subject, such as a supplier, a site or a contract, is run against a set of requirements and returns one finding per requirement, with its evidence attached.
DataStruct AI says each finding points to the page, section or spreadsheet row it came from. Located evidence, citations that could not be confirmed and merely related material are kept apart rather than blended.
When nothing supports a requirement, the finding says so. The product records no compliance verdict of its own.
Every finding goes to a named person who confirms it, rejects it or marks it for follow-up. Decisions are attributed and logged.
PDFs, including scans, plus Word, Excel, CSV, HTML and plain text, searched by keyword and by meaning so evidence can be found even when the document does not use the requirement's wording.
April 2026
Work on the platform begins.
September 2026
The evidence-review workflow becomes available to founding customers, who are onboarded personally.
September 2026
A rebuilt site sets out what the product does, what it does not do yet, and how customer data is handled.
A plain guide to the ventures in MSECO's portfolio, what each one does, how the similarly named ones differ, and who runs what.
Five rules for writing product pages that earn trust, with examples from ventures in the portfolio.
