Implementing GenAI: Maximizing business impact while mitigating risks

Generative Artificial Intelligence (GenAI) continues to make waves across industries — and insurance is no exception. Already, insurers are implementing GenAI into the customer service environment and other business areas, yielding impressive results. While experimentation is crucial to gaining a competitive edge, AI governance must be a high priority to mitigate the potential risks of GenAI.

Integrated GenAI use cases in the insurance sector

Insurers can use GenAI to elevate different use cases, from streamlining critical business functions to optimizing financial performance. AI and GenAI, when used with other rules-based technologies and data integrations, promise to improve both the effectiveness and efficiency of claims and underwriting processes and decision behaviors. For example, digital ingestion solutions using Large Language Models (LLMs) bring greater precision in data extraction and improved document analysis for risk assessment and claim evaluation. These solutions integrate with underwriting and claims technology platforms to achieve higher levels of automation and enhanced decisioning.

Claim fraud detection is an area ripe for change. By combining the power of GenAI tools like ChatGPT or Microsoft Copilot with semantic search and AgenticAI tooling with new sources of data and real-time claim experience benchmarking, greater insights can be unlocked early in the claim investigation process to determine claim fraud potential and mobilize special investigations units teams.

How GenAI enriches customer service

One of the early functions to adopt GenAI and see tremendous impact is customer service – an area long overdue for an AI-driven makeover. By embedding GenAI into call centers, mobile applications and self-service portals, insurers can deliver a more personalized and intuitive customer experience, boosting satisfaction and retention. Unlike traditional chatbots, GenAI-powered agents are aware of customer intentions and engagement patterns, allowing them to provide individualized and context-sensitive information.

GenAI-enabled customer service agents have the potential to simplify what is often perceived as a complex buying process, replacing it with a more satisfying and effective customer journey. Ultimately, these GenAI-fueled customer engagement solutions will empower insurers to move past the conventional data-entry-oriented buying processes and embrace more conversational, needs-identification interactions.

Central to the success of GenAI in the customer service environment and across the wider insurance value chain is Agentic AI, a type of AI that uses sophisticated reasoning and iterative planning to perform complex sequences of tasks without human supervision. In a customer service use case, Agentic AI working with other rules and data platform technologies could offer customized policy quotes based on a customer's risk profile and preferences benchmarked against other similar customers.

The importance of AI governance

The customer service environment is ideal for insurers to test and learn with Agentic AI. However, like any other emerging technologies, Agentic AI and GenAI present insurance companies with just as many challenges to overcome as new opportunities to explore. Insurers should be aware that GenAI can pose risk such as reputational damage, financial loss, compliance challenges and even litigation. In a scenario where a GenAI model drives an improper claim resolution (such as denial of valid insurance coverage), the insurer could be at risk of both litigation and reputational harm.

As insurers implement GenAI to enrich customer service and beyond, they must address data accuracy, privacy and security as well as LLM traceability and explainability through clear enterprise policies and governance. These policies must be prepared to evolve as the technology advances and regulations mature. A risk assessment framework must be leveraged to guide enterprise policies and guardrails. Employees must be trained on these policies and procedures should be established for partner compliance as well.

Inevitably, these AI and GenAI policies will become the focus for audits – both internal and external – and the enterprise should be prepared for levels of accountability to customers, regulators and shareholders even beyond what the industry has experienced with cybersecurity.

The balance of risk and return

The promise of higher levels of efficiency and effectiveness with AI and GenAI is compelling to insurers. At the same time, the new areas of risk potential introduced with these technologies cannot be overstated and present an especially complex challenge for an industry in the business of risk yet inherently risk averse.

Business and IT will need to team in ways not previously experienced given the complexity and pace of change of these technologies and the need to rethink information architecture, process automation and decision architecture. This teaming extends beyond use case and solution development to establishing an integrated approach to risk management and governance across the enterprise level and throughout the business ecosystem.