Controlled AI workflows for recurring administrative work

Measure the starting situation before you automate anything

Record the same process definitions the calculator uses. For each number you note whether it is counted, estimated or unknown.

The short answer

The baseline records where a process stands today: volume, time per case, lead time, errors and exceptions. Do this before you automate anything, otherwise you can prove nothing afterwards. Free, no account, everything runs in your browser and nothing is stored.

What you fill in

The baseline asks for a process name, a measurement period, the number of cases per week and the active minutes per ordinary case. Those four are required, as is the source for the count. Waiting time, rework minutes, the share of cases with rework, exceptions, expected review minutes, backlog, peak load, quality threshold, owner and measurement date are optional, but they are what make the result usable. For every figure you choose yourself whether it is counted, estimated or unknown.

What the tool assumes

The baseline calculates in your browser and stores nothing. It uses the same process definitions as the calculator, so the figures fit there without translation. The handover runs through the URL hash and therefore stays local. The tool does not verify your figures: an estimate stays an estimate, even when the result looks precise.

What comes next

Take the baseline into the calculator and see whether the process is still worth it after review, rework and recurring costs. If you are not yet sure the process suits a workflow, run the process check. If you would like me to look at it, See what you get for free.

Next step

Calculate with the same definitions

The handover to the calculator happens via the URL hash and stays local.

Open the calculator