In Q1 2023, I was working on a project to bring down the costs for one of our cloud services.
It boiled down to a couple of things.
- Get data.
- Make decisions.
We had to collect and work with a lot of data using Google spreadsheets.
That’s where AI tools (ChatGPT) came to my help.
I could describe what I wanted from the data and got pointed guidance on how to do it.
In the process I learned VLOOKUP, pivot tables, many other functions that I had not used before.
It was quick and it was mostly right. And I tweaked what I got to suit my specific needs.
I didn’t share my data with the tool because it was sensitive data and I didn’t trust AI to do the math.
It was more like you tell me what to do and I will check and then do it. Zero trust.
It was still fast. It was accurate. It was useful.
Could I have done it without AI?
Sure. But it would have been more (search) effort, taken more time, and less fun.
This experience was yet another example of the usage of AI which followed this pattern:
- I’m not an expert at something (spreadsheets in this case)
- I know what I want (semantics) and can describe it fairly accurately
- I want to know how to do it (syntax) in that domain
- I want concise output that I can try out quickly and verify for myself
- I don’t want to share sensitive data directly with the tool