Document intelligence, inside the perimeter

Ask an institution where its documents live and you will get a confident answer. Ask what is inside them and the answer gets slower.

Onboarding packs, trade and payment files, statements, contracts, correspondence — this is the raw material of regulated work, and most of it is still moved by people. The systems around it can store a file, version it and audit who opened it. Far fewer can tell you what it says.

Why this is still unsolved

It is not for lack of tooling. Document AI is abundant. The obstacle for institutions in the Kingdom is rarely capability — it is where the processing happens. A model that reads a customer's onboarding pack is reading personal data, and sending that content to a third-party endpoint turns a document-handling problem into a residency and consent problem.

The result is a familiar stall: a promising proof of concept, then a review that asks where the data goes, and a programme that quietly stops.

What we are doing about it

IOTA has partnered with Nutrient to bring document intelligence into our Digital Engineering practice.

Nutrient builds the components applications use to work with documents — viewing, editing, annotation and conversion — along with document intelligence that extracts and classifies content rather than simply displaying it. The property that matters most for our clients is deployment: those components run inside the client's own environment. Document content does not have to leave the perimeter to be processed.

That single characteristic is what moves the conversation from interesting to permissible. It keeps NDMO expectations on classification and residency intact, and it keeps PDPL obligations attached to the data rather than negotiated away for convenience.

Where it lands first

• Onboarding and servicing journeys that are document-heavy and still largely manual
• Legacy document viewers embedded in core applications that have become difficult to maintain
• Structured data that has to be recovered from files arriving as scans and PDFs

None of this replaces the integration work underneath it. Documents still have to reach the right system, under the right controls, at the right time. But once they do, there is now something useful to be done with what is inside them — without the content ever leaving the building.

If document handling is a bottleneck in your estate, we are happy to look at it with you.