AI has been part of quantitative investing for decades. Machine learning, natural language processing and alternative data have long helped investors identify patterns, extract signals and process information at a scale impossible manually. What is new is the rise of agentic AI, where systems are capable of pursuing multi‑step goals, maintaining context and selecting tools autonomously rather than simply responding to prompts. This development raises a different question: instead of asking only whether AI can forecast returns or uncover new signals, asset managers increasingly need to ask where the investment edge actually resides.

From model access to workflow advantage

Early debates around AI in investing focused heavily on model capability. Could AI generate alpha? Could it automate research? Could it reveal previously hidden relationships in data? These questions still matter, but as more firms gain access to similar foundation models, model access alone is unlikely to sustain an edge.

When powerful models can be licensed on broadly similar terms, access becomes less distinctive. The durable edge is more likely to sit in the workflow surrounding the model: how research questions are framed, how data is mapped, how validation failures are recorded, how constraints are applied and where human accountability remains.

You can now read the full whitepaper at the link below