get genai to help with fdd

Do you find clients pressure you to get that FDD red flag report out quicker than you’d like? Well, today’s your lucky day, because GenAI will do a lot of the initial groundwork for you!

This is the prompt I used:

“Please write me a red flag due diligence report, I would like tables detailing the financial position, including variance analysis, plus the top 10 risks the business faces.”

From experience, it’s best to be specific with your prompt. I.e. if I asked just “please write me a red flag due diligence report” you may get a report which is text heavy, rather than getting the required tables, and without structure.

The more specific the prompt the better! You could add further context on the business, on your role, and on the scope of work to create even better results.

Here’s the result from the prompt above:

Sections of the report are downloadable, the most useful sections are the data sections, where you can extract the data and manipulate it in the way you desire in Sheets, or Excel.

You will need to sense check the findings, however, if you put this in a brand Word doc, you will have a nice headstart on your due diligence piece.

Building on the red flag FDD report, you can use GenAI to analyse spreadsheets for you. I uploaded an Excel file and asked it to identify employees whose salaries haven’t increased in the last 12 months, as well as those with a rise of over 10%. This helps highlight staff who may need a raise post-completion or those potentially earning above market rates.

It's also worth noting that AI generated my payroll data. CoPilot is integrated into Microsoft applications, while Gemini is available in Google tools like Sheets and Docs. Below are screenshots of the dataset build prompt (please excuse the spelling!), the dataset itself, and GenAI’s response to my FDD query.

This last example proves the versatility of GenAI. I would argue that most pieces of analysis, and report writing can be assisted by AI. That said, the you need to make sure you are comfortable with the information it produces, and the references used, given there’s not much scope for error!