Broadridge is set to go live with its new algo co-pilot offering in the next few weeks, aimed at helping the buy-side route their orders better.
George Rosenberger
Named NYFIX Algo Co-Pilot, the new product uses AI to build a picture of fils in the market to better inform traders when they look to select an algorithmic strategy, all in a bid to reduce the implicit cost of trading. Broadridge has partnered with Babelfish.com to power the AI element and Google for the cloud requirements.
“This is the first time we’re really seeing AI in practice in trading. Everyone’s talking about use cases but here’s a scenario where it actually works,” says George Rosenberger, general manager of NYFIX, Broadridge, speaking to The TRADE at the FIX EMEA Trading Conference 2025.
“Implicit costs are 85% of the cost of trading. Our goal is to reduce that implicit cost of trading by using these recommendations based on our understanding of how brokers source liquidity from each strategy. We’re using AI to look at all of the dark fils that are going off in the market.”
The co-pilot segments algo offerings in liquidity buckets based on the correlation of how specific securities trade. Within each bucket, it ranks the broker algos based on historical near time performance.
“Think of it like sonar for ships. It takes physical things to figure out a location,” explains Rosenberger.
“It’s the same with dark pools. You can’t see the liquidity but when there is a fil we know the broker and the dark pool it came from. We use AI to match that information up with prints on the tape. With 90% precision we’re able to say what fil came from which dark pool so we know how much liquidity is sitting there.”
The new product will initially float beside the buy-side OMS, however Broadridge plans to integrate it like a blotter in the future – when a trader clicks on a strategy it’ll automatically launch the algo ticket in the OMS.
Unlike an algo wheel or switching engine, the offering will only inform traders as opposed to making a recommendation or automatically routing to a particular broker.
The offering will also monitor changes in the market and update information for the trader accordingly so they can change strategies if necessary.
“Algo wheels are prescriptive and have static rules. This [NYFIX Algo Co-Pilot] allows the buy-side to dynamically make changes based on what is going on in the market,” says Rosenberger.
“We are not making the recommendation on which broker to route an algo to, we are just informing to the buy-side of the choices that they have – this is the optimal strategy for the name they’re trading but ultimately it’s the buy-side directing the order.”