Banks and insurers deploy AI agents to fight fraud and process applications, with plans for new roles to supervise the AI

Core Insights - Financial institutions are rapidly adopting AI agents for customer-facing processes, transforming interactions in banking and insurance sectors [2][3] - AI agents are projected to deliver up to $450 billion in economic value by 2028, indicating significant opportunities for the financial services industry [3] - A majority of firms are in the ideation or pilot stage of AI agent deployment, with only 10% having implemented them at scale [6] Adoption and Implementation - Key processes for AI deployment in banks include customer service (75%), fraud detection (64%), loan processing (61%), and customer onboarding (59%) [2] - Insurers prioritize customer service (70%), underwriting (68%), claims processing (65%), and onboarding (59%) [2] - 33% of banks are developing proprietary AI agents in-house, while nearly half are creating new roles to supervise these agents [3][8] Benefits and Opportunities - Executives believe AI agents can enhance efficiency in customer onboarding, KYC, loan processing, and claims management, with benefits including real-time decision-making (96%), improved accuracy (91%), and faster turnaround times (89%) [7] - 92% of executives see AI agents facilitating expansion into new geographies without heavy infrastructure costs [9] - 79% believe cloud-native AI agents can enable dynamic pricing and offers, enhancing revenue potential [9] Investment Trends - Nearly two-in-three executives allocate up to 40% of their generative AI budget to agent technologies, with expectations of increasing spending on AI agent solutions by up to 60% by 2028 [10] - A growing number of firms (25%) are considering a service-as-a-software model to manage AI costs effectively [12] Challenges to Adoption - Executives identify a skills gap (92%) and regulatory compliance burdens (96%) as major obstacles to AI adoption [11] - High implementation costs are also a barrier, prompting firms to explore new consumption and monetization models for AI [12]