Although most insurance companies have deployed artificial intelligence (AI) in some capacity, most customers in the U.S. have not yet seen the benefits, according to a recent report by Genpact and AWS. The report, titled “Harness the Winds of Change: How to Scale AI and Build Trust in Insurance,” evaluates the industry’s view of AI through the lenses of 200 industry senior executives and 1,000 customers.
AI: Widely Adopted, With Caution
“Traditionally change-averse, the insurance sector is fast realizing that it needs to embrace AI to help navigate today’s volatile landscape and align with customers’ expectations for rapid service,” according to the report. The research shows that AI is widely adopted, use-case-led, and fragmented.
“Most insurance executives (69% globally and 77% in the U.S.) say their companies have deployed the technology in some capacity,” states the report. “But 31% are still experimenting through pilots and proofs of concept without deployment.” The pace of adoption and approach to it vary significantly, “driven by each company’s willingness to innovate, readiness to adapt, and appetite for risk.”
The state of AI adoption is broken down into leaders, fast followers, and laggards in the report. The leaders are defined as those deploying AI applications company-wide, which represents 13% of respondents. Fast followers are those deploying AI in multiple departments (26.5%) or deploying it in one to two departments (29%). Laggards are those developing AI pilots and POCs with none deployed yet (20.5%); evaluating AI use cases with none developed yet (10.5%); and those with AI activity at all yet (0.5%). All in all, AI is overwhelmingly being adopted, or at the very least, being considered.
Deployment: Use-Case Led and Fragmented
Eighty-five percent of AI use cases are identified and driven by business units in insurance companies, according to the report. “While overall direction, budgets, and priorities are largely steered by senior leadership, business units are deciding which use cases to prioritize and how to implement AI effectively.”
The ‘where’ of AI deployment is crucial, however, according to the report. “We see disjointed implementation shaped by the autonomy of business units and geographic regions. This fragmentation not only impacts KPIs and governance—driven by diverse regulations and performance measures across locations—but also influences how and where insurance companies execute and scale their AI strategies, adding complexity to achieving cohesive, enterprise-wide impact.”
Where AI Is Being Deployed
“Data or analytics (57%), claims (46%), software engineering (44%), customer support (41%), and underwriting (31%) top the list of areas where insurance companies are deploying AI,” the report states. “Over the next three years, data or analytics will remain the leader, while customer support, IT, and new business are set to gain ground.” Forty-eight percent of consumers report challenges in claims navigation, indicating that AI investments promise notable improvements in experience and efficiency.
Most executives surveyed (62%) believe AI’s competitive edge lies in high-volume tasks, such as claims processing and customer service, according to the report. However, focusing on productivity as the ultimate goal poses a significant risk, the report warns. “Insurers and intermediaries need to go beyond routine automation and unlock growth and data monetization by addressing complex decisions that require deep research, technical expertise, and contextual understanding,” which is where AI’s transformative power shines.
For claims specifically, “AI can anticipate needs, mitigate risks, and elevate customer interactions. Done correctly, it can help with consistent, transparent decisions without fatigue, delivering reliable, high-quality results time and again,” states the report.
Agentic AI
“Agentic AI presents even greater possibilities,” according to the report. “Imagine systems that can autonomously investigate claims, analyze internal and external data, and present actionable fraud assessments to investigators. It can also refine risk thresholds, adapt to market conditions in real time, and reduce fraud false positives through feedback-based learning.” The result is shorter cycles, enhanced accuracy, and a sharper focus on strategy.
Scaling AI
The biggest barrier when it comes to adopting AI is not technology, according to the report—it is, in fact, governance, risk management, change adoption, and data quality.
Governance and Oversight
The top challenge for insurance companies is governance and oversight (49%). “Data privacy (62%) and regulatory differences across jurisdictions (42%) are regulatory concerns that exacerbate the problem, creating a landscape where a one-size-fits-all strategy simply doesn’t work, “the report states. “Over two-thirds (69%) of respondents report that their companies have an AI governance committee or council, most of which are centralized. The most effective governing bodies have leaders from risk, IT, finance, and insurance-specific functions collaborating to establish charters that address regulatory compliance, risk mitigation, and operational alignment.”
Adopting AI at scale is not just a technical upgrade, but also “an organizational transformation requiring bold leadership and strategic change management,” the report emphasizes. Significantly, the researchers found that “change management is where insurance executives are least confident in making decisions about AI, highlighting the necessity of a focused and structured approach.”
The study recommends building for “scale and trust” with the following steps: strengthening data foundations; balancing governance and risk with precision; prioritizing business outcomes; designing for scale and value from the start; empowering people and managing change; and building a partnership ecosystem for growth and innovation. Researchers encourage laggards to build a strong foundation; fast followers to scale strategically; and for leaders to push the boundaries of innovation.
AI and the Customer Experience
“Despite insurance companies embracing AI, only 36% of U.S. customers…report improvements in their digital experiences with insurers over the past two years,” according to the customer experience companion study. “This may reflect the current limited scope of AI deployment, as insurance companies focus on operational efficiencies rather than customer-facing impacts.”
Some of the reasons for the lack of consumer satisfaction include the fact that companies are not yet ready to allow AI to interact with their customers and require human approval for any customer-facing decisions, the report notes. Furthermore, there is “a crucial trust gap holding insurers back. Only half of customers (50%) trust insurers to provide accurate, personalized quotes…AI, when used responsibly, holds immense potential to help insurers close this gap with greater precision and speed” by delivering quicker and more accurate outcomes in claims and risk modeling.
Nearly half of consumer respondents (46%) noted that they are open to AI if it delivers tangible benefits, such as spending up claims processing, according to the report. When it comes to generational preferences, “Millennials and Gen Z are more likely to choose AI-driven insurers, with 32% preferring AI for policy quotes, 28% for claims, and 28% for customer service—outpacing Gen X and boomers in each area,” according to the report.
Customers still value human interaction, especially during crisis situations, the report finds. “While 41% of insurance executive respondents say their companies use AI to boost customer support, over half of customers (59%) still expect access to live agents, especially during crises. Only 10% are comfortable relying solely on AI-driven chatbots when humans aren’t available.” Therefore, there must be a balance between human touch and AI efficiency.
Uncertain ROI and Poor Data Quality
“Uncertain ROI [(return on investment)] tops the list of significant pain points regarding AI for insurance companies…and misalignment between AI ROI frameworks and business goals hinders deployment at scale.” In fact, just over 35% of insurers indicated that they currently have the right KPIs (key performance indicators) to measure AI.
Furthermore, “Insurance companies need robust data frameworks and capabilities to apply governance frameworks and build AI tools…companies relying on aging legacy systems must address their technical debt first. Successful deployment of AI needs a modern, integrated data architecture. Data privacy, unclean data, and integration challenges often prevent scaling.” This understanding of data’s role in AI is reflected in budgeting priorities, with 51% of executives indicating they are investing in data quality over the next one to three years and 43% focused on data integration and interoperability.
AI Fluency: A Top Priority
The talent gap faced by the insurance industry continues, and when it comes to AI, only 2% of insurance executives indicate that nearly all their team members are AI fluent. Meanwhile, 69% of respondents say that either very few or some of their employees are AI fluent. “To bridge this divide, over half (53%) plan to build in-house AI expertise within the next six months to two years…among laggards, this figure rises to 65%,” states the report. “AI fluency spans two streams: employees who bridge tech and business, understand AI’s potential, and develop use cases; [and] everyday users comfortable integrating AI into their workflows.”
Executives will need to prioritize finding talent with both insurance expertise and AI fluency, the report emphasizes, which may prove challenging. “It’s not just about algorithms but about mastering policy structures, regulations, and customer behaviors to turn tech into actionable, impactful solutions.”
The top five ways insurers are building internal AI expertise include training for existing employees/upskilling (46%); hiring specialized talent (44%); creating a dedicated AI team (39%); fostering a culture of continuous learning (39%); and creating a center of excellence (CoE) (38%). “Investing in AI talent pays off far beyond technical implementation,” states the report. “AI-fluent teams make faster decisions using insights for underwriting, fraud detection, and pricing. They reduce errors by automating repetitive tasks, spot risks like model drift or bias early, and enable smoother enterprise-wide transformations.”
The Importance of Upskilling Employees
Although only 46% of insurance companies are focused on upskilling current employees, 81% of leaders versus only 33% of laggards are dedicated to upskilling by combining internal training with external partnerships. “Personalized training is driving progress, with companies crafting role-specific learning paths and enabling basic AI literacy across the board,” explains the report. “Leading organizations are already preparing for agentic AI, teaching teams to collaborate with intelligent agents, delegate tasks, and manage decisions.”