Centralised AI models propel financial institutions' success

Financial sectors see major benefits from centralised AI strategies, though future shifts are anticipated.

Financial institutions are increasingly turning to centralised general AI (Gen AI) models to drive success, according to Violet Chung, Senior Partner at McKinsey & Company. These centralised models have shown significant advantages in scalability and resource allocation across various business units, despite the broader trend of decentralisation in other areas.

"Across industries, we've found that a high degree of centralisation works best for Gen AI operating models," Chung explained. Central oversight helps prevent pilot AI use cases from becoming isolated within silos, which can make scaling efforts much more challenging. This centralised approach has been particularly effective in the financial services sector, where institutions with centrally-led Gen AI models are reaping substantial rewards.

Chung outlines four primary archetypes for adopting Gen AI in financial institutions: highly centralised, centrally led but business unit executed, business led but centrally supported, and highly decentralised. Despite these variations, some degree of centralization is maintained to manage execution centres and ensure efficient resource allocation.

One of the critical benefits of centralization, Chung noted, is the enhancement of capability building within organisations. "In a very centralised setting, an AI Center of Excellence can support many stakeholders, executing crucial functions and managing processes and resources, including external data," she says. This support is vital for fostering competence and innovation as organisations further integrate AI into their operations.

However, implementing centralised Gen AI models is not without its challenges. Institutions must have a clear, top-down aligned strategy, adequate talent with the necessary skill sets, and adaptable technology and data architectures. "The most important, and often underestimated, element is a new operating model that accommodates these advanced technologies," Chung adds. This involves reorganising processes, addressing skill set needs, and making strategic decisions that enhance organisational comfort with new AI capabilities.

Looking to the future, Chung advises organisations aiming to implement centralised AI models to start with a clear identification of where the most value can be found. "Aligning the organisation's efforts from the top down is crucial for achieving significant breakthroughs and impacts," she states. This strategic alignment should be coupled with realistic assessments of return on investment and necessary expenditures to support these advanced technologies.