Generative AI in Insurance Claims Faces a Scale Problem
While 78% of property and casualty insurers have adopted generative AI, only 4% have scaled it meaningfully across their claims operations, according to a new Bain & Company survey of 160 global insurers—revealing a significant gap between experimental adoption and transformative implementation.
Generative AI deployment in P&C claims remains largely experimental, Bain found. Most insurers are using the technology to handle specific tasks such as fraud detection, document summarization, and customer communication. However, this piecemeal approach falls short of what’s possible when AI is embedded throughout the entire claims process, the report said.
Insurers taking an end-to-end approach experience dramatically different results. Those redesigning their full claims operations around AI capabilities have achieved a 35% productivity boost, reduced settlement cycles, and cut homeowners’ claims processing times in half, according to Bain.
Barriers to Full Implementation
Despite these compelling benefits, only 27% of insurers are pursuing comprehensive claims transformation, Bain said.
Three primary obstacles are holding the industry back: data security and privacy risks, insufficient in-house expertise, and concerns about accuracy. These challenges are significant enough that most insurers remain in pilot or limited-use stages rather than committing to organization-wide AI redesign, according to Bain.
View the a survey infographic here. &

