By Bas Kooijman is the CEO and Asset Manager of DHF Capital S.A
Allocators face a torrent of signals, including economic data, policy shifts, correlation breaks, and liquidity changes. In this regard, AI can help in detecting regime changes earlier and stress-testing portfolios across many plausible paths.
Models trained on decades of cross-asset data can identify when relationships between different asset classes are shifting. Policy tone is another important factor, and text sentiment analysis can provide valuable insight into the impact on markets, making it easier to translate signals into allocation views. Scenario analysis is evolving, too. Managers can test portfolios under different scenarios and risk stress, providing a deeper understanding of potential growth and risks.
However, the operational payoff is the leverage of human judgment. AI filters the deluge of data and signals while people make the decision. Pairing advanced models with rigorous human governance can provide a significant edge. The approach includes incorporating guardrails against overfitting, challenging models, and maintaining a human sign-off on asset reallocations. Used this way, AI becomes a decision infrastructure and a performance enabler, not a black box. Such systems help turn complexity into clarity faster without abandoning discipline.