Openbox AI – Just describe the model you want

Unless you’ve been having a total digital detox, you’ll have seen articles about AI being used in all sorts of ways. And one of those ways – since the current AI wave is all about language – is to write things. Need some code for your website? Ask ChatGPT. Need a summary of some complex document? There’s an AI app for that. Need to write a blog…wait, I promise you this blog was written by me, without any AI help.

So why not use it to write financial models? People have been sceptical about this because financial models are complex beasts, with lots of interlocking parts. And they’re probably right that the combination of AI + Excel is not enough to come up with a full financial model on its own.

However, Openbox operates at a higher level of model structure than Excel. And it turns out that the combination of AI and Openbox can start automating your financial model build process, as this video outlines. You can say to our Openbox assistant “I have a solar project in Spain, with two tranches of debt. Model construction monthly and operations quarterly” and in a couple of minutes, out will come a complete model, in Excel, just like one you might have taken hours to make yourself.

Seeing this can bring out some extreme reactions! The first one is that AI is going to replace financial modellers. Nice for clickbait articles, but not true, unless you think that typing standard formulas into Excel is the only thing financial modellers do. What about understanding client requirements? What about presenting results clearly? What about… the list goes on. Frankly, I’m always delighted to delegate the boring parts of my job!

If AI is not a replacement for financial modellers, what is it? It seems to work best if you treat it as a helpful but junior assistant that knows how to translate what you say you want into a precise description that Openbox can then use. And that feels completely natural. It’s no coincidence that when you see computers in science fiction, people don’t usually have to operate them through buttons or a user interface that they need to learn. They just tell the computer what they want and the computer does it,  “…playing middleman between the rigidity of a computer system and the anarchy of an organic one”.

That to me is the real benefit here. The gap between “I know what I need a model for” and getting a solid draft of that model just shrank to almost zero.


If you’d like to test it out, you can sign up for our private beta waitlist here.

Share:

More Posts