This week, I continued using the ChatGPT Work feature to refine my prototype and get it ready to submit for peer review.

I focused on the recommendation screens, which are the part of the tool where the user sees the financing pathways suggested for them and the reasoning behind each one.

When I asked for changes, it wanted to make too many at once. I started guiding it in a more narrow way, being specific about what I wanted changed and what I wanted left alone. That made it easier to see what was actually different and to keep the parts I already had working.

This is something I have noticed a few times now. In Week 3 the agent I used suggested extra sections for my website that I had not asked for, and in Week 4 I wrote about AI offering more than what I needed. Narrowing what I ask for has made a difference.