Aflac is a global leader in supplemental health and life insurance, providing financial protection to over 50 million policyholders in the U.S. and Japan. 

In 2023, Aflac reported an annual revenue of $18.7 billion. With approximately 12,785 employees worldwide, Aflac continues to drive innovation in cancer and medical insurance.

Although Aflac’s total investment in AI is unknown, we found ample evidence of its applications in enhancing operational security and the customer service experience. 

In this article, we focus on two key use cases of data-driven AI models at Aflac:

• Improving cybersecurity capabilities: Leveraging predictive analytics to detect threats and reduce false positives, improving data protection and consumer trust.
• Optimizing the customer service experience: Using machine learning and natural language processing (NLP) to automate claim processing and email classification, leading to cost and time savings.

Improving Cybersecurity Capabilities 

Phishing attacks and identity theft are on the rise as generative AI elevates the sophistication and speed of cyberattacks, according to a 2024 study by the Lloyds Banking Group. The FBI’s Internet Crime Complaint Center 2023 crime report finds that accounts of phishing attacks more than doubled between 2019 and 2023.  

Adversaries can now collect vast troves of consumer data and exploit vulnerabilities faster than ever. However, a CrowdStrike blog post asserts that two core issues previously paralyzed Aflac’s ability to contain breaches: 

• False Positives: Aflac relied on a legacy managed security service provider (MSSP) but struggled to keep pace with alerts, of which 99% were false positives. 
• Fragmented Security Stack: Aflac adopted various top-notch cybersecurity solutions since no single vendor sufficed. The company spent significant capital, time, and resources to integrate and maintain those tools.

“Speed is always a challenge in cyber because the faster that adversaries can exploit vulnerabilities, the faster you have to be at identifying them,” DJ Goldsworthy, Aflac’s Vice President of Security Operations and Threat Management, said in a video interview with SiliconAngle.

According to use case documentation from CrowdStrike, Aflac needed to utilize AI in its defense system in 2018, developing capabilities that detect and curb attacks at the same speed as adversaries. The case study claims Aflac partnered with Crowdstrike to consolidate its solutions and promote cybersecurity resilience.

The CrowdStrike briefing also discloses that Aflac deployed the AI-driven CrowdStrike Falcon cybersecurity platform, starting with CrowdStrike Falcon® Insight XDR, which provides endpoint detection and response. Crowdstrike claims the model employed a generative AI security analyst that accelerates incident detection, reducing detection and mitigation times from hours to minutes. 

Below is a graphic of CrowdStrike’s generative AI model built on AWS: 

Screenshot from an Amazon Web Services Partner Network (APN) blog post on CrowdStrike’s Charlotte AI. (Source: Amazon Web Services)  

CrowdStrike claims their AI models facilitated greater visibility of Aflac’s digital environments, enabling security teams to identify previously invisible misconfigurations and vulnerabilities. Armed with AI-driven insights, Aflac implemented new access restrictions and automated micro-segmentation capabilities, making it more difficult for cyber threats to gain unrestricted access. 

CrowdStrike documentation also notes that Aflac also eliminated 15-point security tools by AI to consolidate on the Falcon Platform. The singular solution enabled more agile threat-response capabilities and cut capital expenditures significantly. 

Finally, the documentation claims that the Falcon Platform and enhanced threat intelligence delivered:

• Fewer False Positives: False reports dropped by 20x, enabling quicker and more effective responses to real threats
• Greater Agility in Threat Detection: The time it takes to detect threats dropped from hours to minutes 
• Improved Staff Availability: Efficiency enabled by automation allowed over half of Aflac’s security team to focus on more complex tasks
• Increased Consumer Trust: Improved threat mitigation capabilities strengthen consumer trust and allow customers to focus on recovery without concerns about data breaches 

 Optimizing the Customer Service Experience

Aflac serves over 50 million people globally, many of whom hold multiple claims. A Pega case study corroborates that managing this scale using traditional processes led to various challenges: 

• High Call Volumes: Long wait times and high abandonment rates negatively impacted customer satisfaction.
• Disjointed Systems: Legacy applications and disparate manual processes hindered staff’s ability to respond to customers, requiring extensive training for new hires.
• Email Overload: Thousands of weekly email inquiries required manual classification and responses, delaying resolution times.

The Capgemini Research Institute’s World Life Insurance Report 2025 quantifies that across the industry, 25% of customers are frustrated by long wait times. Legacy technology and disjointed systems remain a significant barrier to improving customer service for 52% of insurers, according to the same report and Capgemini’s 2024 Global Insurance Executives survey.

Aflac joined forces with Pega to automate several parts of the customer service experience to reconcile such challenges.

According to the Pega briefing, Aflac began using the Pega Platform’s AI analytics to integrate 25 disjointed back-end systems and gain an accessible 360-degree view of each customer. The capability ultimately reduced the need for agents to juggle between several applications and improved clientele communication. 

Another case study from Coforge also maintains that Aflac used Pega’s platform to automate email processing. Manual classification of 3,000 weekly emails and processing across seven inboxes previously hindered timely consumer service. The automation capability leverages natural language processing (NLP) to understand email context and sentiment. 

Coforge contends that the Pega platform was able to: 

• Automate Processing: Classifying and assigning emails to appropriate cases.
• Streamline Resolution: Forming intelligent and quick automated responses to simple inquiries that don’t require agent interaction.
• Expedite Manual Responses: Generating suggestions and insightful input for manual resolution by staff.

Similarly, over 75% of inbound chats and inquiries were handled by virtual AI agents equipped with NLP, according to the Pega briefing documentation. 

In an interview with the MIT Sloan Management Review, Aflac’s Chief Information Officer Sheila Anderson revealed that many categories of insurance claims were also  automated using “straight-through processing.” Such claims don’t require additional proof of loss and only need attestation from the policyholder to serve as proof. 

AI and machine learning models thus rely on claim form answers and attestations from policyholders. The workflow integrates AI to classify claims, automate adjudication, and streamline payouts, as depicted in the figure below:

Screenshot from an Intelliarts blog post on machine learning in Insurance workflows via vehicle damage recognition capabilities. (Source: Intelliarts)  

“Automate the easy and service the complex,” Keith Farley, Aflac’s Senior Vice President of Individual Voluntary Benefits, said in a conversation with CMSWire.

In the same interview with MIT Sloan Management Review, Sheila Anderson disclosed that nearly 46% of Aflac’s claims were automated, which enabled staff to dedicate time to high-value, emotionally sensitive, and complex requests. 

Pega further claims the following business results for Aflac:

• Greater customer satisfaction and retention
• 33% reduction in handling time for claim requests
• 65% reduction in effort for client authentication
• More than $4 million saved through automated chatbots