Why AI Success Doesn’t Travel Across the Insurance Value Chain

By Samarth Bhutani, Manager, Solution Consulting at Material 

AI is delivering measurable gains across underwriting, claims, service and operations, yet few insurers have translated those wins into enterprise-wide impact. Recent findings from the Evident AI Index for Insurance suggest that while AI is improving productivity across the insurance industry, enterprise-wide business impact remains limited.  

Across the insurance value chain, successful use cases are emerging in almost every function. Yet despite growing investment across the industry, evidence of enterprise-wide impact remains limited. Insurers are moving toward operating models where decisions, policyholder journeys and risk continuously inform one another, but few have reached that level of maturity.  

The harder question is whether the intelligence, efficiencies and decisions generated in one part of the business can improve outcomes in the next. The ability to carry those gains across the value chain increasingly separates incremental AI adoption from genuine operating model transformation. 

Where the AI Value Is Already Showing Up 

Agentic workflows are increasingly demonstrating the ability to accelerate end-to-end business processes, not just individual tasks.  

• BCG found that AI agents handling insurance claims from intake through payout have reduced claim handling times by up to 40 percent in some deployments while improving customer satisfaction scores by as much as 15 points.  

• McKinsey documented Aviva’s deployment of more than 80 AI models across its claims organization, reducing liability assessment times for complex cases by 23 days, improving claim-routing accuracy by 30 percent and reducing customer complaints by 65 percent. 

These are significant business results that remain confined to the functions where they were created. 

Beyond claims, AI is helping insurers improve sales conversion, accelerate underwriting decisions, increase premium growth and streamline policy administration. Research across the sector suggests leading insurers are realizing meaningful gains in both productivity and customer outcomes through targeted AI deployment. 

Yet the value created by these initiatives does not always show up across the broader organization. A claim contains intelligence about risk. An underwriting decision contains intelligence about pricing. A policy change contains intelligence about customer behavior. When those signals fail to move consistently across the value chain, their impact remains local.  

Enterprise performance depends on how effectively operational intelligence moves across the organization, not simply on how many AI initiatives exist. Understanding why those gains stall requires looking beyond individual use cases and examining the operating model that connects them.  

What’s Preventing Enterprise Value 

Even when AI initiatives deliver measurable gains, gaps in three business drivers (coordination, data and continuity) often prevent those gains from translating into enterprise-wide impact. The table below highlights where those gaps emerge and how they limit enterprise value. 

Viewed together, they reveal where value stalls between successful initiatives and enterprise outcomes. 

What a Connected Operating Model Looks Like 

In an AI-enabled enterprise, a connected operating model creates an intelligence-first enterprise loop, where every interaction, decision and outcome strengthens the next. 

This requires three conditions to exist simultaneously: 1) Operational Alignment, which connects teams around shared business outcomes; 2) Connected Intelligence, which ensures operational signals inform decisions across the value chain; and 3) Enterprise Scaling, which embeds successful ways of working across the organization. 

Material’s work with an insurtech client demonstrates how this model operates in practice. The insurer’s claims process relied heavily on manual vehicle inspections, creating delays between assessment, claims handling and settlement. Vehicle damage assessments took several days to complete, slowing decisions throughout the workflow.  

The redesigned process transformed a series of disconnected activities into a connected workflow, where intelligence generated at one stage remained available to every stage that followed. 

The process is embedded into operations, so intelligence generated during the first interaction becomes immediately available to subsequent decisions, routine cases no longer compete for specialist attention and human expertise is concentrated where it creates the greatest value.  

From Investing in New Tools to Maximizing Current Capabilities 

AI’s biggest economic gains come from redesigning how work actually gets done. Two recent industry estimates show how much that discipline is worth within a single domain. 

• Rewiring entire domains such as claims or underwriting is helping insurers achieve 10 to 15 percent higher premium growth and 20 to 40 percent lower customer-onboarding costs. 

• Integrating multimodal AI and advanced analytics into fraud detection can reduce related costs by 20 to 40 percent, depending on the insurer and the maturity of its systems. 

Each of these findings measures what a single domain can do on its own. Material’s analysis examines what happens when those sources of value reinforce one another. In one modeled scenario, stronger coordination among sales, underwriting and claims creates value beyond what any of those functions could generate independently. That shifts the board’s conversation from funding the next capability to maximizing the value of the capabilities already in place. 

Before the next AI investment gets approved, three questions are worth asking. 

1. Where do gains stop at the edge of one function instead of reaching the next?
2. Where does operational intelligence stop moving through the value chain?  
3. Which parts of the operating model determine whether a successful initiative becomes enterprise capability?

Material’s financial services experts work with leading insurers to uncover opportunities for greater enterprise value across existing capabilities. Get in touch to continue the conversation.