It’s easy to understand intuitively that financial markets are predictive to some extent. In other words, there are things we can learn through data, and other things that are fairly opaque and mysterious, at least until you apply some pretty advanced analysis.

But I think that a lot of people would associate finance, in general, with learning. People want to understand methodologies, and how finance works – they want to be able to make better decisions by being more informed, not just about what markets are doing or how to budget money, but how to forecast and make predictions and plan for the future, which is kind of what budgeting is when you think about it.

In fact, investing is the same, in a way. Obviously, it works better if you have a long-term plan.

Using AI for Finance

So given all that, it seems like finance is a perfect application for the products of large language models, where these analytical machines are so highly cognitive that they can make predictions and teach us more about how to handle money.

I recently saw a range of presentations by some great young career professionals who are developing real-world use cases for artificial intelligence.

Two of them had to do with finance, but they had very different angles applied.

Event Prediction

The first one was a predictive engine from Aroosh Krishna that analyzed the Kalshi exchange – although it has an Indian-sounding name, the exchange was created in the U.S., fairly recently.

I wasn’t familiar with Kalshi, but when I read up on it, the platform gets at least 1 million users routinely, and you can bet on everything from natural disasters and elections to things like what Trump is going to say in a certain time frame.

They call it event-based investment.

In any case, a project called AnalyseKalshi uses sentiment analysis to help predict outcomes. There are two APIs involved, according to a flow chart showing how this engine works.

Krishna notes that part of the goal is to level the playing field between hedge funds and what he calls “casual investors.”

I found this to be an interesting use of AI, although applied to a market that seems to lack a certain seriousness.

In other words, our tools can make highly predictive recommendations about certain kinds of outcomes based on the capricious emotions of, in some cases, rather unstable people, or more serious and substantial predictions … driven by highly technical pattern observations that LLMs can help us with.

Teen Budgeting

The second presentation had to do with creating personal finance tools for teenagers or young adults.

Presenter Viren Kedia noted that the app is meant to help teens to be able to do things like invest, and budget, and understand taxes.

Notes on program research on the presentation banner show that in two days, teens got a 30% increase in literacy quiz scores.

That’s important because the quiz scores show you, as a user, what you’re learning, and help with the benchmarking that the app brings to the table.

Users, Kedia adds, like three major things about the application: free version, a good interactive chatbot, and video functionality.

Teens learn the difference between a budget and financial plan, and setting financial goals.

It’s all part of educating this segment of the user base in better money management, and yes, in a way, predictive capabilities.

Can these tools achieve what we need them to? The plans look actionable, and there’s a clear value proposition to both of them. I’ll continue to bring some of the most interesting and compelling projects to the blog, so stay tuned.