Only 36% of Major Finance Players Achieve Full AI Integration — Guess the Main Obstacle

According to a study conducted by a leading data analytics and digital operations company, financial services and insurance giants are investing heavily in AI, with 91% having launched AI POCs in the past year. However, a mere 36% have successfully integrated AI across their business operations.

The research, based on responses from 158 influential decision-makers in the finance and insurance sector, highlights the hurdles faced in achieving comprehensive AI implementation.

While the majority (55%) of firms have limited AI applications within their organizations, only 36% have embraced company-wide AI initiatives. The focus areas for integrated AI include marketing and business development (47%), risk management and fraud detection (43%), internal operations such as claims management (42%), and back-office billing and payments processing (37%).

The primary obstacle is the growth of data silos, hindering enterprise-wide AI integration for 74% of firms using AI for limited functions. These data silos are reported to be within specific business functions (33%) or shared among some functions while isolated in others (41%).

Meanwhile, generative AI projects raise concerns for 54% of respondents. The top issues include algorithms operating beyond intended parameters (44%), potential new regulations (43%), safeguarding customer data (42%), and the risk of biased decision-making (42%).

For those 54% who have already implemented generative AI, the focus is primarily on product development (93%), customer care/experience (82%), human resources (82%), and corporate strategy (75%). Future adoption of the same is anticipated in regulatory compliance (52%), risk management (52%), and corporate strategy (52%).

Vivek Jetley, Executive VP and Head of EXL Analytics (the organization that conducted the study) emphasized that despite the industry's recognition of AI's potential, the primary roadblock remains data integration.

Siloed data, often trapped in legacy systems, impedes the realization of AI's full potential, posing a significant challenge for businesses aiming to transition from conceptual projects to fully integrated solutions.