The problem

The Saudi Industrial Development Fund ran internal processes the way large public institutions usually do: documents moving between departments, approvals collected in sequence, and a person responsible for carrying each item to the next desk. Every one of those handoffs is a place where work stalls, where the current status is whatever somebody remembers, and where the audit trail has to be reconstructed after the fact from mailboxes. The cost is rarely a single dramatic failure; it is a steady tax of chasing, re-sending and re-explaining that nobody measures because it never appears on a system report.

My role

Contract engagement through ASGA Tech, delivering enterprise government solutions inside the fund's existing Microsoft estate. I worked on the document management and workflow automation systems — the routing logic, the document handling around it, and the data layer that made both of them reportable.

Architecture

• SharePoint 2019 as the platform, because it was already the institution's document system. Replacing it was never the goal — automation that forces people to leave the tools they already use does not get adopted, it gets worked around, and then you have two processes instead of one.
• Document management built on that same platform, so a document has one authoritative location with its own permissions and version history rather than a filename with a date on the end travelling through attachments.
• Approval routing implemented as explicit workflows instead of email chains: each step knows who it is waiting on, what the item is, and where it goes on approval or on rejection.
• C# for the logic the built-in workflow engine could not express — the conditional routing, validations and integrations that separate a form from an actual process.
• SQL Server behind it, so the state of every document and every approval is queryable data rather than something buried in an inbox, which is also what makes the process measurable in the first place.

Technologies

SharePoint 2019, C#, SQL Server.

Outcome

Manual operational steps dropped by 40%. Inside a government fund that is not primarily a speed story, it is a consistency and accountability story: steps that used to depend on someone remembering to forward a document now happen the same way every time, and the state of any request is a lookup rather than a phone call. Working under public-sector constraints also fixed a rule I still design to — automation has to be auditable, not just fast. If a system cannot show who approved what and when, it has moved the problem rather than solved it.