Leading global asset managers are deploying artificial intelligence to analyse their own vast internal archives of research, communications, and historical decisions in a bid to discover unique investment signals. With AI commoditising access to public and alternative data, firms like BlackRock and Balyasny Asset Management are now focusing on proprietary information inaccessible to competitors.
The shift was highlighted at the Future Alpha conference in New York, where executives detailed how large language models (LLMs) are being used to structure and interrogate decades of unstructured internal data. This internal mining is seen as the next frontier for generating alpha, or market-beating returns, as traditional data advantages erode.
The Commoditisation of Public Data
Jacob Bowers, Vice President of Quantitative Research at BlackRock, stated on a conference panel that AI has made once-cutting-edge, publicly accessible data "commoditised." With the world's largest asset manager overseeing $14 trillion, BlackRock has already directed AI agents to scrutinise past communications between investment professionals and old opportunity reports. "Some of the best unstructured data you have is internal," Bowers said, noting AI's proficiency at organising such information.
This strategy capitalises on a long-understood potential. A 2019 Opimas report suggested funds might eventually monetise their proprietary data. Robert Frey, a former Renaissance Technologies managing director, previously told Business Insider that his old firm's "massive data library" accrued over decades was its key advantage.
Structured Processes for AI Analysis
At the $33 billion hedge fund Balyasny Asset Management, the process is more structured. Andrew Gelfand, a quant focused on alpha capture, explained that the firm mandates analysts to input all research and notes into a central portal. This creates a rich, text-based dataset for AI systems to analyse for predictive signals. Gelfand confirmed that while previous attempts to monetise this unstructured data were challenging, recent AI advances have made the effort "much more fruitful."
The efficacy of this approach hinges on the quality of the internal data—the insights and reasoning of top-tier investors. Mike Daylamani, who leads a team blending fundamental and systematic investing at Engineers Gate, emphasised this point at the conference: "You need the feedstock to be high quality," he said, referring to the data used to build quantitative models.
The Human Element in a Data-Driven Race
Despite the technological arms race, industry leaders stress that investing remains a fundamentally human, creative endeavour. Daylamani added, "At the end of the day, this is a creative endeavor." The goal is not to replace human judgement but to augment it by uncovering latent patterns and connections within a firm's own intellectual history that would be impossible for any individual to process manually.
As AI capabilities expand, the focus for quantitative and fundamental teams alike is shifting from merely acquiring more external data to better leveraging the unique, high-quality informational assets already contained within their own organisations' walls.