The Rise of Generative AI in Insurance

March 13, 2024

By Stan Smith

The global market for artificial intelligence (AI) in insurance is predicted to reach nearly $80 billion by 2032, according to Precedence Research. This growth is being driven by the increased adoption of AI within insurance companies, enhancing their operational efficiency, risk management, and customer engagement.

Despite widespread integration of AI in the industry today, its full potential is in its infancy. While AI has proven its capabilities in automating and excelling at insurance-specific tasks such as assessing risk, expediting claims and performing complex analysis, many insurers have yet to harness these efficiencies despite the challenges they face in workforce shortages and intensifying competition.

In this article, I’ll explore the potential of combining generative AI (GenAI) with traditional AI as a catalyst for achieving more profound and impactful transformations.

Enhancing Efficiency and Effectiveness

Efficiency and effectiveness drive the insurance industry. Efficiency involves achieving more with fewer resources and less time, at a lower expense. Effectiveness entails making better decisions and driving better outcomes. The integration of generative AI and traditional AI enhances both efficiency and effectiveness.

Traditional AI excels at analyzing vast datasets, including historical policy and claims data, offering a comprehensive evaluation of underwriting and claims risks. Meanwhile, using large language models, GenAI can be trained to perform tasks like summarizing notes, writing emails and providing insightful guidance. 

This combination streamlines insurance underwriting and claims processes, enabling insurers to make better decisions about risk, increasing policy pricing accuracy and enhancing claims outcomes.

Empowering the Next Best Action with Generative AI

Generative AI has the potential to significantly reduce insurance claim costs and duration by performing time-consuming tasks and guiding adjusters toward optimal actions. It can analyze a vast amount of data to provide actionable recommendations.

Imagine an insurer handling a worker’s compensation claim for an injured employee. Traditionally, the process would involve reviewing medical records, consulting healthcare providers and manually assessing the worker’s condition to determine the appropriate course of action. This can lead to delays, prolonged worker absence, and higher claims costs.

Leveraging traditional and generative AI, the adjuster inputs data such as medical reports, diagnostic test results, adjusters’ notes and job requirements. Traditional AI algorithms can leverage available data, analyzing past similar cases and monitoring the recovery of an injured worker continuously. Based on this analysis, GenAI can provide appropriate recommendations for the best next actions, like:

• Optimized next steps: Suggest an independent medical exam or other steps already demonstrated effective for similar claims.
• Modified work assignments: Recommend appropriate modified duty assignments, ensuring a smoother transition back to work.
• Rehab plans: Suggest tailored rehabilitation or physical therapy programs, expediting recovery and potentially reducing claim duration.
• Cost-cutting strategies: Propose cost-effective measures, such as utilizing in-network medical providers.
• Communication recommendations: Offer guidance on the best communication practices between the injured worker and healthcare providers, ensuring a smooth flow of information.

Improving Confidence Through Explainability 

A key concern in AI adoption is the concept of “explainability” or the system’s ability to explain how it makes decisions. Traditional AI models can seem like “black boxes,” leaving professionals perplexed. GenAI addresses this by providing interactive decision support, explaining results in plain language, and even engaging in conversations. For example, if a medical term appears in a claim, insurers can ask open-ended questions and engage with GenAI to obtain a clear, easy-to-understand explanation, enhancing their confidence by clarifying the factors impacting risk predictions. GenAI helps users comprehend the reasoning behind the model’s conclusions, playing an important role in establishing trust and accountability, essential in the insurance industry.

Challenges Adopting AI in Insurance

The collaboration of traditional and GenAI holds immense promise for insurers. However, there are challenges that must be addressed.

Data Quality: Assuming availability in a usable format, real-world data often contains inaccuracies and inconsistencies, posing a significant challenge in maintaining clean, reliable data for AI models. Even when collaborating with an external vendor, compatibility and preprocessing of data for model consumption remain crucial. Identifying and rectifying issues like missing or inconsistent data is essential.

Addressing Biases in Machine Learning: AI systems learn from data, which, if not representative of the entire population of a covered group, can cause AI models to perpetuate or even amplify particular biases. Imbalanced training data can lead to inaccurate or discriminatory predictions. To mitigate biases, the training data must be diverse and representative of the population. Including a broad range of demographics, backgrounds, and experiences helps prevent the model from learning and perpetuating biases.

Data Quantity and Variability: Insufficient data access is also a challenge. Successful machine learning relies on substantial datasets that capture diverse inputs and outcomes within specific business contexts. For instance, when training an AI model to predict workers’ compensation claims, factors like injury types, treatments, and locations significantly impact outcomes. Relying on a narrow dataset might lead to inaccurate predictions across regions, injury types, or severities. 

Cultural Acceptance: Building trust in AI is essential for successful integration and collaboration between employees and AI systems. This includes offering training to familiarize employees with AI – what it is, how it works, how it complements their roles, its benefits, and addressing any misconceptions or fears. Model explainability helps, as does emphasizing how AI is simply another tool to amplify their work, rather than to replace human capabilities. Professionals can leave routine, repetitive tasks to the AI systems, and increasingly focus on more strategic, value-added tasks.

The Future of Generative AI in Insurance

As we look to the future of the insurance industry, it’s clear that integrating generative AI and traditional AI is key to unlocking AI’s full potential.

The combination promises increased efficiency and effectiveness, enhanced decision-making, and optimal utilization of the existing workforce. Traditional AI excels at predicting risk, policy pricing, and projecting claim reserve requirements. GenAI’s natural language capabilities will accelerate insurers’ ability to make sense of large amounts of unstructured data, enabling them to make more informed decisions quickly and accurately.

The synergy between these two AI approaches will help insurers achieve a better return on risk, stay competitive in an increasingly challenging market, and deliver excellent client service.