AI in regulated industries: BofA and S&P on the huge gains to be had, and the risks of rushing in

At the recent Fortune AIQ Summit in New York, top executives from Bank of America and S&P Global shared critical insights into the integration of artificial intelligence within highly regulated financial sectors. Hari Gopalkrishnan, Chief Technology and Information Officer at Bank of America, and Sally Moore, Chief Client Officer at S&P Global, emphasized that while AI offers transformative potential, the most common mistake organizations make is rushing to implement it as a primary solution without first evaluating simpler, deterministic alternatives. Both institutions are currently navigating the delicate balance between aggressive technological investment and the stringent risk management required in global finance.

Bank of America is significantly scaling its AI efforts, with plans to double its AI budget next year. CEO Brian Moynihan noted that 140 current AI use cases have already generated $800 million in benefits against $400 million in costs. The bank utilizes a rigorous 16-pillar risk review process for every project, covering concerns from bias to intellectual property. Meanwhile, S&P Global, leveraging its 2018 acquisition of Kensho, has repositioned its business structure to prioritize data accuracy and auditability. Sally Moore highlighted a case where S&P helped a tier-one bank increase process accuracy from 60% to 98% while accelerating production timelines by six times, proving that targeted AI agents can deliver measurable, high-impact results.

Both companies possess deep historical roots in the financial industry, which informs their cautious approach to innovation. Bank of America has been utilizing AI for over a decade, starting with fraud detection models, and its virtual assistant, Erica, has successfully managed over 3.6 billion transactions. S&P Global, a 160-year-old institution, views data as the fundamental currency of AI. By creating a dedicated chief client office and specialized labs, S&P ensures that its AI deployments are not just experimental but are deeply integrated into the workflows of its 60,000 global clients, ensuring that every model remains traceable and reliable.

The ramifications of these technological shifts are profound, affecting everything from workforce productivity to systemic financial stability. Gopalkrishnan warned that the industry must remain vigilant regarding security, noting that as AI models improve at identifying software vulnerabilities, companies must prioritize secure development regardless of their AI adoption level. Furthermore, the failure of a smaller institution could erode public trust in the entire financial system. Consequently, both firms advocate for a 'human-in-the-loop' approach, ensuring that AI acts as an assistive tool rather than an autonomous replacement for professional judgment.

Looking ahead, the consensus among these industry leaders is that the future of AI in finance lies in 'embedded intelligence' and the strategic use of assistive agents. Rather than chasing total autonomy, firms are focusing on enhancing human capabilities to serve clients more effectively. As control infrastructure matures, the ability to anticipate client needs and deliver precise, data-driven solutions will become the primary differentiator in an increasingly commoditized financial landscape. The path forward requires a disciplined, model-agnostic strategy that prioritizes the right tool for the specific task over the mere pursuit of the latest technological trend.