Quantitative investment teams are using Artificial Intelligence and advances in natural language processing and large language models to extract and translate vast amounts of company text – filings, earnings calls, news and reports – into machine-readable data. The article notes that roughly 50 to 60 percent of global data is text-based and that rises to 70 to 80 percent for business-relevant information. Northern Trust Asset Management highlights the scale: about 8,300 public companies in the MSCI ACWI IMI index have issued roughly 230,000 filings in 21 languages, a volume that is now tractable thanks to cheaper computing power and modern models.
That textual intelligence is vectorized into numerical representations that feed network construction. Networks are composed of nodes, the individual companies, and edges, the connections between them. Edges can be weighted by metrics such as transaction value or percentage importance. Examples given include supply chain networks, where suppliers and manufacturers form nodes and transactions define edges, and technology networks built from patent-ownership data. These networks reveal a company’s ecosystem, peer group, supply-chain exposures and how information flows across related firms.
The practical investment implication is to use network-derived signals to augment proven style factors. The article explains how combining classic momentum signals with network-based momentum can help identify more persistent, stock-specific drivers and reduce crash risk associated with sector- or country-level reversals. Network analysis enables more granular risk budgeting and more deliberate portfolio construction, offering a potential source of incremental alpha for quantitative strategies. Guido Baltussen, head of quantitative strategies international at Northern Trust Asset Management, frames network signals as a new lens grounded in economic theory, enabled by technology, and validated by empirical work.
Looking ahead, the piece argues that the challenge for quantitative investors is discernment in a world awash with data. Networks provide context that can make models more risk-efficient and improve outcomes by separating transient noise from economically meaningful relationships. In that way, the use of Artificial Intelligence to map company connections opens new frontiers for alpha generation while supporting more resilient portfolio decisions.
