- Tech for Finance
- ChatGPT Code Interpreter + Excel, 3 way models, and creating AI assistants
ChatGPT Code Interpreter + Excel, 3 way models, and creating AI assistants
1-2-3 Newsletter - Read Time - 4 Mins
Hello, and welcome to a another edition of the Tech for Finance 1-2-3 Newsletter where we have:
1. Tutorial - ChatGPT Code Interpreter + Excel
2. Expert resources - On 3 ways models & revenue recognition
3. Tech tools - To help you create your own AI assistant, work remotely and Q & A your documents
1/ ChatGPT Code Interpreter + Excel
ChatGPT Code Interpreter is now here (exciting) and it's power is mind boggling.
In this example I'm using dummy data that you can find on this link, and using ChatGPT to manipulate, analyse and make predictions from the data.
IMPORTANT - Make sure your data is anonymised, or at least free of any sensitive information.
Step 1 - Upload your data
You do this by clicking the plus button. It's also good practice to ask ChatGPT whether it understands the file you've uploaded.
"Please read and confirm that you understand the contents of this document."
Step 2 - Re-format the data
What I want to do here is create a new sheet within the spreadsheet that shows a quarterly summary of profit by product.
I can do so using a prompt like:
"Please create a new sheet within the spreadsheet that shows monthly profit by product. Do not combine data from different years. Please ensure they remain separate."
Here's what the downloaded data looks like:
Step 3 - Q & A the Data
Let's do a bit of historical analysis. I'm curious to know our most profitable countries, and whether there's any correlation between volume of units sold, or discount bands.
A prompt like this could work:
"OK, looking at the whole document, what are our most profitable countries? Are there any correlations between volume of units sold and discount bands? If you feel a visualisation will help us understand this, then please choose the most appropriate one."
Step 4 - Further Analysis
Let's go a bit further.
"Looking at the data, does number of units sold always equal higher revenue?"
Step 5 Go with the flow
You might see ChatGPT make recommendation on how to continue to Q & A.
In this example, it's asked whether we want to visualise the correplation with a scatter graph.
I just answered 'Yes' and this is what it produced.
Step 6 - Make predictions
Now comes the REALLY cool part. Using AI to make predictions.
"Thanks. Now using the entire dataset, please predict how how future unit sales are likely to increase or decrease"
With the above, ChatGPT has used a simple model to try and predict future values.
But it doesn't have a large amount of historical data to work with.
It's also come back with some things to note about forecast accuracy (and it also apparently doesn't understand colours).
It also did a LOT of working out before producing the screenshot above.
It you want to see the full thread, click here. Just bear in mind you won't see the images.
The new code interpreter is an incredibly powerful tool. But it's still important to ensure your instructions are clear, especially when manipulating data, as you don't want to have to start again several times.
The speed is also impressive. Especially when re-formatting data, I can see this saving hours of time.
There's a bit of a learning curve, but well worth the investment in time.
2/ Expert Resource
2. This week, I interviewed Em Daigle from Zuora Revenue. It was a great discussion where we covered everything including:
- How to think more strategically
- How to assess revenue tools
- Go-to-market strategy
- Revenue accounting
- AI & Automation
- Escaping Ex-hell
- Much more
For the full podcast interview 'From Bean Counter to Business Partner: The Rev Rec Edition | Ep 029' go to the podcast here, and for the YouTube video go here
3/ Tech Tools
1. MindOS - Create your own AI assistant for professional tasks - learn more here.
2. SoWork - You AI powered remote 'smart' office - learn more here.
3. Afforai - Q & A documents, spreadsheets & websites - learn more here.
IMPORTANT - Always read the data policy for any tools you try. If you’re ever unsure. Make sure your data, and any customer or supplier data is anonymised.
NOTE - I am not affiliated with any of the tools in this e-mail.
So that’s it for this week.
Until Next Time
👨💻Connect with me on LinkedIn
🎧Subscribe to the Podcast
📺Watch on YouTube