Find the friction.
Look for work people repeat. Read the same document. Move the same data. Check the same detail. Ask the people doing it where the time goes.
Set a baseline.
Measure the work before you change it. How long does it take? Where do mistakes happen? Who checks the result? A clear baseline makes a small trial useful.
Build a narrow test.
Choose one task, one group of users, and one way to judge the result. Use real examples. Include the difficult ones. Keep a person in the loop where a mistake would matter.
Decide what comes next.
Keep what works. Change what does not. A useful test may lead to a larger system. It may also show that a simpler tool is enough. Both are good results.