Human-in-the-loop Reasoning
When prompting the AI to create a commission tracker to save time on manual commission calculations each week, I wanted it to be clear and concise. With this tracker, I could save approximately 3 hours per week by reviewing each collected check and calculating commissions manually. I would need the calculation to be accurate, as this was the most important facet for me while I manually calculated the commissions. Upon receiving the tracker, it was truly organized and simple to use. One sheet tracked individual commissions, another gave the totals for each person, and the last sheet provided a summary for the whole pay period. I ended up tweaking the sheet a bit, splitting up the individual commission tracker into quarters. But the formulas in the sheet were easy to use and could be edited if needed. The tracker's interface was user-friendly and clear. The factors that shaped the outcome of this sheet include the language I used to prompt the AI, as well as its Excel knowledge and the formulas it used. I was very clear about what I wanted the sheet to do and what my goal was. Clarity allowed the AI to refine the sheet and gave it a clearer, more concise sense of direction. This allowed the sheet to specifically work for me, rather than being a general sheet for anyone. There were some tweaks I had the AI handle before I went in and made some changes myself, but the AI did much of the heavy lifting, saving me time in my week.