Finance and capital markets are highly regulated industries. What happens when data and AI are added into the mix? Does compliance overhead decrease by automating oversight-related documentation and reporting? Does the potential for crossing regulatory boundaries increase and quicken?

To make sense of how new deployments of data may change the capital market space, we look at three regulatory variables shaping the relationship between data and capital markets.

Data collection, storage, and deployment

Whereas the cost of mistakes is low in a field like e-commerce—you received the wrong product, and you get a new one after contacting customer service—the potential consequences are far higher in more regulated fields like capital markets and medicine.

This forces players in those sectors to track how they collect, move, and store data, in addition to analysis and deployment—compelling them to seek out data management solutions with a history of working in high-compliance spaces. From GDPR to PCI DSS to SOC 2, a comprehensive framework makes market actors solve for compliance through more significant regulatory-focused headcounts and systems.

More individual responsibility

The relationship between financial leaders and data is changing. Where teams used to hire “data stewards” in charge of storing and deciphering data, executives—and individual contributors—face growing pressure to deal with data themselves.

This means that, while compliance-focused teams still exist in full force, data-focused compliance concerns are diffused more extensively across an organization. Ensuring that data is collected in line with legal frameworks, for instance, is something on several employees’ checklists, rather than on a “data steward’s” alone. That reality—and inconvenience—may push leaders to automate more of these distributed compliance tasks.

Black boxes

Finance-focused applications of data and artificial intelligence require transparency. Black box solutions can infringe upon consumers’ rights, and obfuscate the reasoning behind an application’s approval or rejection.

We should expect data solutions to include clear reasoning behind their outputs, perhaps adding another AI layer—a generative AI layer, for instance—as the linguistic and qualitative component.