You said the office team is thinking through what they'd want automated. Phase 2 starts with one day of interviews, then I disappear to make a plan, come back with ideas you can pick from, and build whatever you point me at. The goal is to ship every low-hanging win we can find in the time I'm given.
I want to start by understanding who these people are. What they care about. What they are good at. What they wish they were doing more of, and what is grinding on them at work. If something outside of work comes up that makes weeks harder than they need to be, that is a blessing too. It goes into the calculus of where the leverage is.
The frame is closer to ethnography than user research. The best automation candidates surface when you understand the human, not just the workflow. A 45-minute conversation per person, recorded with consent so I am listening instead of writing. All four in a single day on-site.
One day off-site to turn what I heard into a written proposal. Per-person notes, cross-role patterns, a ranked candidate list. Each candidate gets a one-paragraph description, an effort estimate, an impact estimate, and a why-this-one. Bias toward low-hanging wins that ship fast and matter to a real person on the team.
Sit-down with you and the team. Walk through what I heard, what I am proposing, and which one I think is the right starting pilot. You and the team push back, edit, point me at the favorite target. The pick is yours.
Same delivery rhythm as Phase 1. Working code, shipped fast, on real S&S workflows. I pull from your team for context whenever the build needs it. When one win is shipped, we move to the next target on the list. The list keeps growing as the team sees what is possible.
I won't know the real candidates until after the interviews. To show the shape of the eventual document, here are three illustrative branches based on the categories you mentioned.
Timekeeping is what surfaces
THEN the build menu might look likeInventory is what surfaces
THEN the build menu might look likeQuoting and purchasing is what surfaces
THEN the build menu might look likeThese three branches are illustrative. The actual list comes from the interviews. An operation this size has countless ways to find efficiencies, and the more we learn and build the more surface area we will uncover.