August 21, 2025 AI and the Fight Against Property and Casualty Fraud

Multimodal technologies can sniff out and help thwart a multibillion-dollar drain on the industry.

By Kedar Kamalapurkar, Managing Director and a leader in the insurance sector claims practice, Deloitte
Michelle Canaan, Insurance Research Leader, Deloitte Center for Financial Services
Namrata Sharma, Insurance Research Executive Manager, Deloitte Center for Financial Services

An estimated 10 percent of property and casualty (P&C) insurance claims are fraudulent, resulting in a US$122 billion loss annually, or 40 percent of all fraud losses within the insurance industry.ⁱ After tax evasion, insurance fraud is the second-most costly white-collar crime in the United States. ^{ii iii}

Raising premiums to make up the loss is not an attractive answer. Not only are premiums already under enough pressure from inflation, but those rates already reflect the cost of fraud: According to the FBI, insurance fraud adds US$400 to US$700 to an average American family’s premiums every year.^iv The Coalition Against Insurance Fraud reports that 78 percent of US consumers are concerned about it.^v

Generative AI (GenAI) enters the fray

Equipped with the new detection models and advanced data analytics GenAI makes possible, insurers can spot anomalies more quickly while also freeing up human investigators to focus on more complex fraudulent cases across the claims lifecycle. In addition, AI-fueled multimodal technologies, which use AI to process and integrate data from multiple sources such as text, images, audio, video, and sensor data, can generate more detailed and accurate insights than earlier systems. Laws that govern the use of these tools vary by jurisdiction, but many insurers have the opportunity to use them to enhance their capabilities to detect and prevent fraud.

In a recent Deloitte survey of insurance executives, 35 percent of respondents chose fraud detection as one of the top five areas for developing or implementing GenAI applications over the next 12 months.^vi

A race to greater sophistication

Anti-fraud efforts have evolved alongside the fraud attempts that make them necessary. In particular, the increase in digital communications and transactions that followed the COVID-19 pandemic created new opportunities for fraudsters as well as a flurry of innovative solutions to combat them.^vii Fraud detection technology has become a rapidly growing industry, estimated to multiply eight-fold, from US$4 billion in 2023 to US$32 billion by 2032.^viii At the same time, pressure from the National Association of Insurance Commissioners is pushing insurers to implement effective and advanced fraud detection systems.^ix

The technologies available to answer that call are many and varied. Techniques such as automated business rules, embedded AI and machine learning methods, text mining, anomaly detection, and network link analysis can score millions of claims in real time. Multimodal detection helps identify patterns and anomalies and enhance the investigative process by reducing false positives, increasing detection rates of fraudulent claims, and saving on investigative costs. None of these tools works alone; each is meant to work under effective human oversight.

• Text analytics: Natural language processing analyzes textual data of claims forms, emails, and social media posts to identify keywords and entities, and can flag suspicious language or inconsistent details for further investigation.
• Audio-image-video analysis: Speech recognition and sentiment analysis can examine customer calls for signs of duress, while photo analytics can uncover irregularities in meta data, manipulation, and repeated use. Causation analytics can test whether alleged injuries were consistent with a reported accident. Video analytics can verify damage, authenticate images, and highlight signs of tampering or staging.
• Geospatial analysis: Satellite images and comprehensive 3D drone footage can verify the extent and location of damage that may not be clearly visible in physical inspection.
• Internet of Things data: Real-time surveillance devices like vehicle telematics and smart home sensors can reconstruct accidents or verify damage. Security cameras can help detect fraudulent or staged activities.
• Simulation models: Using a virtual environment to replicate the behavior of medical providers, repair shops, and others can identify deviations from standard industry practices, detect overbilling or unnecessary services, and reveal collusion.

Better fraud detection results in potential savings

Using current methods, detection rates are 20% to 40% for soft fraud (which includes opportunistic cases such as exaggerating a claim or omitting information) and 40% to 80%  for hard fraud (deliberate, premeditated deceit).^x Insurers that integrate multimodal capabilities using AI and advanced analytics could generate potential savings of 20% to 40%, depending on the implementation, type of insurance, and sophistication of fraud detection systems. ^{xi xii}

Based on estimates from The Coalition Against Insurance Fraud and The National Insurance Crime Bureau, Deloitte expects that implementing AI-driven technologies across the claims lifecycle, while integrating real-time analysis from multiple modalities, could reduce fraudulent claims and potentially save P&C insurers between US$80 billion and US$160 billion by 2032.

A new partnership between humans and technology

P&C insurance involves a great deal of complexity and high-volume data, which carriers need to process quickly. That means the potential for cost savings and efficiency improvements is great as well.^xiii

Special investigative units to detect and mitigate fraud are not new to the industry. But today’s pressures are. GenAI has the potential not to replace human professionals, but to bring new speed and focus to their work. Insurers that pair sophisticated technology with human enablement can put the “art” in artificial intelligence, detect more claims fraud more quickly, and potentially save billions of dollars for themselves and their policyholders.

About the Authors:

Kedar Kamalapurkar is Managing director and a leader in the insurance sector claims practice at Deloitte Consulting LLP.  He has nearly 15 years of experience in claims operations, including as a claims adjuster.  He has led claims transformations from strategy to execution for many of the major insurance carriers in the United States and Europe. Kedar also holds insurance industry professional designations from CPCU, AIC, API, and AINS.

Michelle Canaan is a Manager in the Deloitte Center for Financial Services in New York. As a subject matter specialist for the insurance industry, Michelle produces thought ware on current and future trends, including strategies and solutions for our clients.

Namrata Sharma is a research manager at the Deloitte Center for Financial Services. She has over 15 years of research and due diligence experience in investment banking, credit analysis, and asset management with major financial services firms, and launched her own B2C startup. Sharma has used her strong data analytics experience to generate actionable insights in mergers and acquisitions, rating transition matrix, portfolio optimization and grass roots–level business execution. As part of Deloitte’s research team, she specially enjoys crafting data-driven graphical roadmaps/frameworks for future looking thoughtware.

ⁱ Covers fraud from property, auto, workers compensation. Coalition Against Insurance Fraud, National Insurance Crime Bureau,  “Insurance Fraud Statistics”, January 03, 2025, Analysis by Deloitte Center for Financial Services

ⁱⁱ Defined as intentional deception in the insurance process, from policy purchase to claims settlement.

ⁱⁱⁱ National Insurance Crime Bureau,  “Insurance Fraud Statistics”, January 03, 2025

^iv National Insurance Crime Bureau, “NICB and Agero Join Forces to Combat Insurance Fraud”, June 12, 2024

^v Coalition Against Insurance Fraud, “Insurance Fraud Statistics”, January 03, 2025

^vi DI to replace this placeholder citation with the link of the Deloitte Insurance AI Readiness publication – To be published in April ahead of the predictions

^vii Satish Lalchand, Val Srinivas, Brendan Maggiore, Joshua Henderson, “FSI Predictions: Deepfake banking and AI fraud risk”, Deloitte Insights, May 29, 2024

^viii Global Market Insights, “Insurance Fraud Detection Market Size, Growth Trends ”, May 2024 How AI Is Helping Insurance Companies Fight Against Fraud – Business Insider

^ix Timothy Cercelle, Irena Gecas-McCarthy, Jim Eckenrode, “2025 Insurance Regulatory Outlook, Deloitte Insights, February 21, 2025

^x Discussion with Kedar Kamalapurkar, Deloitte Insurance Claims leader, Dec 13, 2023, Analysis by Deloitte Center for Financial Services.

^xi Coalition Against Insurance Fraud, “Insurance Fraud Statistics”, January 03, 2025, Discussion with Kedar Kamalapurkar, Deloitte Insurance Claims leader, Dec 13, 2023, Analysis by Deloitte Center for Financial Services.

^xii Combining data from various modalities, such as text, images, audio, and video, could help identify patterns and anomalies and enhance the investigative process and reduce fraud claims.

^xiii National Association of Insurance Commissioners, “Insurance Fraud Statistics”, January 03, 2025

Subscribe to Industry Today