Why We Build AI Trading Tools Instead of a Hedge Fund

“We tried AI and it didn’t work and if it does work, why don’t you guys trade with it?”

My co-founder and I have heard versions of this from prospective customers (and, to a degree, from investors). Last week we wrote about the first half of that objection (Why there is still a gap in AI Trading). This week is about the second: why we’re building tools instead of a hedge fund.

Picks-and-shovels

Most people know Levi’s, the blue jeans company. Probably not as many are familiar with their history and how they started as a small fabric and dry goods store in San Francisco during California’s gold rush and grew to a multinational, multibillion success story. Levi Strauss didn’t go after the golden nuggets like many of his peers but instead sold cloth and tools to the miners (this early success is what later led to the highly durable working pants and blue jeans as we know them). 

You know the Levi’s brand, but how many of the gold miners can you name? That’s the "picks-and-shovels” model in a nutshell. It tried and tested, can be hugely successful, and that is what we are going after. 

Why picks & shovels make sense for financial services

• Level the playing field. Big firms have used proprietary tools to stay ahead of smaller players in trading for years. This market dynamic means a lot of potential firms are willing to subscribe to tools that can level the playing field.
• Current technology landscape. The new AI architectures that are powering the modern LLMs can now be applied to finance providing robust short-term predictions. Building such computational tools is a huge challenge for small and large firms. A small VC backed team with the right skills can build what many organizations don’t have the patience, expertise, or resources to accomplish. We talked about this changing landscape and the gap in existing approaches here.
• Less direct competition. Tech stacks are siloed inside funds and guarded as secrets. That creates whitespace for a focused software player to build industry standard tools.
• The market never sits still. Regimes shift, data shifts, microstructure shifts. Many in this space are always looking to try new approaches and there are few barriers to entry. 

What we offer and how we deliver

Here are the specfiic tools we provide: intraday price-trend predictions. Specifically, for any point during the trading day, we estimate the probability that the price will be higher or lower by end-of-day. Our stack and business model are built for easy consumption and straightforward pricing.

Under the hood: We license exchange data, train finance-native transformer models on our backend, and run them against the live feed. Predictions are delivered in real time via a clean API. The delivery and commercial model mirror a familiar data-feed workflow, so teams can plug us in without changing their stack.

‍