Balyasny Asset Management’s head of applied AI is spearheading an initiative to streamline analyst tasks through the implementation of AI technology. Samantha Lee from Business Insider reports that Balyasny is in the process of developing a range of AI-driven bots that can efficiently perform the duties typically handled by senior analysts. These bots are specifically crafted to enhance the productivity of investment teams by automating menial tasks, ultimately aiming to elevate data analysis capabilities to unprecedented levels on Wall Street.
The asset management firm is actively constructing an AI-powered counterpart to a senior analyst by amalgamating various AI tools, as disclosed by the hedge fund’s head of applied AI to Business Insider. This exclusive story is accessible solely to Business Insider subscribers. Balyasny has been enhancing its customized version of ChatGPT, known as BAM ChatGPT, which was rolled out to all employees in June 2023. By integrating BAM ChatGPT, developed using OpenAI’s API and hosted on Microsoft Azure, with internal and external datasets, Balyasny endeavors to tap into ten diverse sources, including transcripts, sales data, sell-side commentaries, and broker research.
The primary objective is to establish a network of bots capable of proactively delivering pertinent information—such as real-time updates on relevant companies or discrepancies in company disclosures—to portfolio managers (PMs) and other business units. Charlie Flanagan, the head of applied AI, emphasized the focus on generating proactive insights and anticipatory analysis to stay ahead of the curve. The transition from junior analysts to senior analysts is a key priority for the team, aiming to empower the agents with advanced analytical capabilities.
Hedge funds across Wall Street are eager to leverage the AI trend for operational efficiency gains. Various firms are harnessing AI technologies for diverse purposes, from coding assistance at Man Group to market data analysis at Two Sigma. Bridgewater is even developing an AI investor to execute trades with client funds.
The integration of AI is revolutionizing the workflow of Balyasny’s investment teams, facilitating easier information discovery within extensive datasets and automating labor-intensive tasks. For instance, AI bots developed by the head of research and his team significantly expedited the monthly analysis process for a recurring market event, reducing the timeline from two days to just 30 minutes. Additionally, AI holds great promise in streamlining the analysis of regulatory filings like 10-Ks, enabling automatic identification of discrepancies in disclosures and prompt dissemination of relevant information to analysts.
While the AI agents are not intended to replace analysts entirely, they are projected to alleviate approximately 10% of their workload, transforming analysts into editors rather than creators of analyses and summaries. As Balyasny progresses towards AI integration into daily operations, Flanagan’s team, comprising researchers, data scientists, and engineers from esteemed institutions like Citadel and Goldman Sachs, has expanded to ten members in recent months.
BAM ChatGPT’s approach involves developing specialized “agents” with distinct mandates to handle specific tasks, which are then combined to yield comprehensive outputs. By encouraging analyst and PM teams to construct their own agents using AI building blocks, Balyasny has witnessed early successes, with teams uncovering valuable insights that have led to profitable outcomes. The adoption rate of BAM ChatGPT stands at around 80% among Balyasny’s workforce of approximately 2000 employees.
Furthermore, Balyasny is actively refining the models to enhance their comprehension of numerical data, graphs, and charts—an area where the current OpenAI GPT model exhibits limitations. Flanagan highlighted the importance of improving the models’ ability to interpret information presented in table and visual formats, especially from broker research reports and other documents received in PDF form.
