The spreadsheet is a decision you keep making

Manual work rarely announces itself as a problem. It arrives as a favour someone did once, under pressure, that nobody ever took back out.

Every organization I have worked inside had a spreadsheet holding something important together. Participant records. Disbursement tracking. Who approved what, and when. None of it started as a decision. It started as a workaround, and then the workaround became the process, and then the process became the institution.

The cost is easy to miss because it is paid in small amounts. An hour reconciling two versions of the same list. A morning chasing an approval that lives in someone’s inbox. A week each quarter assembling a report from six sources that do not agree. Nobody logs those hours as a problem, because the work does eventually get done.

At the Kosmos Innovation Center I mapped how the organization actually worked, which is different from how the org chart says it works. Then we rebuilt it as one system covering participant data, dashboards, approvals, disbursement, procurement, and recruitment, with automation layered along the chain. Airtable recognized me as a power user for the build, which was flattering, but the number that mattered was how much time came back to the team.

The hard part is not the tooling. It is that every manual step is protecting something: a judgement someone does not want to write down, an exception the process never handled, a relationship that works because a person is in the middle of it. You have to find out which of those are real before you automate anything, or you build a system people quietly route around.

So I start with the map, not the software. Where does work wait. Who is copying data from one place to another. What does a person do every Monday that a system should have done on Sunday night. Once that is visible, the build is the easy part, and the organization stops renewing a decision nobody remembers making.