I think the sensible answer may be to stop expanding our FR5 bag-to-carton trial. We began with 109 hardware kits. Transferring them is possible; maintaining the presentation for each new bag shape is what keeps growing.
I am the person who has put the time into it, which probably makes me a poor judge of when to stop. Has anyone retained a small useful part of a trial without committing to automate the entire range?
Kept a vision check for one stable product family after a wider trial became all setup and no production. Had to give that family an actual boundary, otherwise every new variant was 'nearly the same' and I was back where I started. For your bags I'd compare the stable subset with the manual job, including who keeps the presentation kit usable when you're away. Could be worth keeping. Could also be a nice demo you learnt from. The hours already spent don't decide which