Analyst To AI

In a previous post, I prompted Copilot to suggest ideas for how I can use AI in my role as a Director of Finance for a nonprofit. Based on the output, I decided to take a suggestion and put it into practice. I am trying to make the link between the “potential” that this technology has been promising and how we can get it to work.

I asked Copilot to prepare a cash flow forecast using an Excel file that I created where I track each days bank transactions from our key operating bank accounts. I do this so that I can 1) see the activity going through the bank and spot anything that I wasn’t expecting 2) check that our accounts have sufficient balances for the days activity 3) move funds between accounts as needed.

The output took about 5-7 minutes to create. It based the forecast and built its assumptions on almost 4 years of historical data.

It was presented in table format for the numbers in the excel spreadsheet with a narrative on the right side. I didn’t prompt this but it was shown in monthly increments, when the file data is recorded daily. I DID prompt it to forecast the remainder of 2026 through the end of 2027, which it did accurately. It also adjusted for seasonality, which again I didn’t prompt, but was impressive.

In keeping true to my mission of learning about and understanding how to use AI and keep me relevant in my role in finance, I thought this was a good exercise to try. I asked AI what it was capable of and it responded with a number of tasks, including cash forecasting. Now, I could take the hype of AI and put it to practical use.

The verdict is that I thought it did a pretty good job of saving me time. The 5 – 7 minutes it took to read the file and convert it into tables and explanations after a couple of more minutes, was a lot less time than what it would’ve taken me at this level of sophistication. It took the data and calculated averages then projected it forward. I also liked that it inserted a table where I could manually adjust assumptions to try different scenarios or to “true up” anything that I would think appropriate. I’m going to spend a little more time with this file to understand how it came up with the output but first glance it looks pretty good.

I’m not a critic and that isn’t the purpose of this blog. I’m only trying to understand this technology and learn how to use it in my role in finance. By journaling it, I’m learning it. I’m also learning what AI can do and its limitations so I can compare it to the hype.

This technology is definitely coming I believe and it’s important that those of us in finance can use it. What I’ve learned so far takes away a little bit of anxiety that I’m going to be replaced by a robot, however I also learned that I have to embrace it to stay relevant. I hope anyone who reads this can benefit from my time and efforts. Also please share any insight or experiences.