How to choose initial RPA pilots in insurance back-office to demonstrate ROI?

[AbigailJ]

I'm an operations manager at a mid-sized insurance firm, and I've been tasked with identifying high-volume, repetitive back-office tasks for potential Robotic Process Automation to reduce errors and free up staff for more complex work. I've started mapping processes in claims data entry and policy renewal notifications, but I'm struggling to quantify the ROI and build a business case that justifies the initial investment in an RPA platform and developer time. For those who have implemented RPA, how did you select your first pilot processes to ensure a quick win and demonstrate value? What were the biggest unexpected challenges in terms of change management with employees or maintaining the bots long-term as underlying systems change?