Chaz Perera is the Co-Founder and CEO of Roots, a company pioneering the use of AI Agents to revolutionize the workplace.
Since bursting onto the scene a few years ago, generative AI has rapidly transformed from an emerging tool to a business imperative across many industries, with insurance being no exception. While many insurers have implemented generative AI for operational efficiencies, some are just scratching the surface of what this technology can do.
Insurance, by nature, embraces innovation. However, insurers are forced to take a deliberative approach to implementing new technologies, and generative AI is no exception. My company, which provides an agentic AI platform for insurers, found in our 2025 report that 82% of respondents view AI as a top strategic initiative. Despite this, only 22% of insurers reported having AI solutions running in production.
The discrepancy between acceptance and adoption is understandable, given the insurance regulatory environment, insurers’ talent and knowledge gaps and other practical constraints. However, it’s important that organizations take steps to keep up with this technological change curve. Otherwise, they risk missing opportunities to build lasting customer value and gain a competitive advantage.
Beyond The Obvious Applications
Many insurers have successfully deployed generative AI for relatively straightforward use cases: automating claims documentation, improving customer service chatbots and extracting information from policy documents. These implementations deliver meaningful operational improvements, but there are additional areas insurers can explore.
Underwriting Intelligence: I’ve observed many carriers use AI primarily for data extraction in underwriting rather than risk analysis. To develop more sophisticated risk assessment frameworks, insurers can use AI to synthesize insights from disparate data sources, such as satellite imagery, social media trends, macroeconomic indicators and climate models. This approach could enable more accurate pricing, creating parametric insurance products and customized coverage options tailored to emerging risks.
Claims Optimization: While a growing number of insurance businesses use generative AI to process standard claims documentation, I’ve noticed few have implemented comprehensive claims optimization systems that use AI. With the right guardrails in place, AI can be used to help detect fraud patterns, recommend optimal settlement strategies, predict litigation probability and provide personalized communications to policyholders.
Product Innovation: Perhaps the most overlooked area for improvement is product development. Insurers can use generative AI to help analyze consumer behavior, market trends and loss data to identify embedded insurance opportunities, coverage gaps and emerging risks that traditional actuarial approaches might miss. Insurers can then use these insights to design products that address evolving customer needs.
Understanding Implementation Barriers
Why are some insurers hesitant to use AI for use cases like these? Several factors are at play here:
1. Risk aversion: Many insurance executives are concerned about the explainability and governance of AI systems, particularly given regulatory scrutiny. This can lead to limitations on AI’s application.
2. Regulatory compliance: Insurance is heavily regulated, with requirements varying across jurisdictions. Many insurers struggle to develop and implement AI solutions that simultaneously deliver innovation while ensuring compliance with evolving regulations around data privacy, model transparency and consumer protection. This regulatory complexity can inhibit the development and deployment of AI applications.
3. Skills gaps: Many insurance organizations lack the technical talent and cross-functional expertise to implement sophisticated AI solutions. Effective implementation requires collaboration among data scientists, underwriters, claims specialists and business strategists.
4. Legacy systems: Aging infrastructure can also create significant integration challenges that complicate AI deployment beyond standalone applications.
5. Cultural resistance: The industry’s traditional approach to decision-making can clash with the iterative, experimental approach required for successful AI implementation.
Charting A Path Forward
To fully capitalize on generative AI’s potential, insurance leaders can consider these approaches:
Start with business problems, and embrace responsible experimentation.
Identify specific use cases where AI can deliver meaningful impact and “quick wins.” Don’t just implement the technology for its own sake. Then, create controlled environments through proof-of-concept exercises so teams can test AI applications with real data while maintaining governance guardrails.
You can build on this by establishing robust AI governance best practices. This can accelerate successive AI implementation decisions and enable future AI innovation.
Streamline future legacy system integrations.
Legacy “green screen” or 1990s systems often lack application programming interfaces, making them difficult to automate using traditional techniques like robotic process automation. Given this, insurers might start by identifying a process that’s difficult to automate and involves a legacy system integration; then, explore how AI could be applied.
For instance, an AI agent could receive a goal like “open a new claim” and apply agentic capabilities to successfully navigate the required system prompts and create a repeatable workflow without having to bring in staff to code a solution. This approach can help insurers gradually adopt AI tools without overhauling their entire infrastructure.
Implement feedback loops.
Deploy AI with processes built around definable key performance indicators to measure AI performance against business objectives. The use of human-in-the-loop systems can ensure continuous expert oversight for improving AI systems’ accuracy and transparency.
Partner strategically.
If you’re working with AI providers, the most important factor to consider is ensuring the vendor can work with your data in your formats to deliver the data outputs to your systems. That ensures they have the ability to tailor AI models specifically to your insurance use cases. And, ensure the provider can scale to support your production volumes.
As I see it, the insurance companies that will be in the best position in the next five to 10 years will be the ones that understand how to use generative AI as a strategic asset for the organization’s core business functions. AI can offer opportunities to introduce fresh perspectives for underwriting, claims, product development and customer experiences—but only for businesses whose leaders adopt it strategically and are prepared to navigate the hurdles.
Forbes Business Council is the foremost growth and networking organization for business owners and leaders. Do I qualify?
Read the full article here
