Allstate Insurance finds AI chatbots more empathetic than human representatives
AI chatter is everywhere today as it’s use becomes more and more mainstream. It’s used in many aspects of insurance, including claims processing, underwriting, pricing, and marketing.
Oh – and maybe as well as being a nerd with numbers, it’s also nicer than you.
The Wall Street Journal has reported that Allstate has found that artificial intelligence, devoid of human emotions, is proving to be more empathetic than many of its own representatives. The insurer has increasingly turned to AI-generated communication in handling customer claims, a move it says has led to clearer, more compassionate interactions.
According to Allstate’s CIO, the company now relies on OpenAI’s GPT models to craft nearly all of its claims-related emails, ensuring they align with company-specific terminology while also reducing the use of industry jargon.
According to Allstate Chief Information Officer Zulfi Jeevanjee, this shift has significantly improved customer interactions.
“When these emails used to go out, even though we had standards and so on, they would include a lot of insurance jargon. They weren’t very empathetic…Claims agents would get frustrated, and so it wasn’t necessarily great communication,” Jeevanjee said in an interview with the WSJ.
With 23,000 representatives handling approximately 50,000 customer communications daily, AI now writes the majority of these messages. Claims agents still review them for accuracy, but they are no longer responsible for drafting the emails.
“The claim agent still looks at them just to make sure they’re accurate, but they’re not writing them anymore,” Jeevanjee added.
Customer service has become a leading area for generative AI adoption as companies seek to streamline interactions while maintaining a personal touch. In the insurance sector, where claims resolution often involves complex exchanges, AI’s ability to maintain consistency and clarity is proving valuable.
Allstate’s PR department has since gone into overdrive, trying to get publishers to remove Jeevanjee’s quotes – it won’t be drawn on whether they are accurate – as it appears they would rather have some more spin-worthy quotes to replace them.
Josh Streets, a senior consultant at the International Customer Management Institute, noted that while Allstate is ahead of many competitors in deploying AI-driven communications, other insurers are quietly exploring similar applications. However, he emphasized that the debate over whether AI or human representatives provide better empathy remains unresolved.
Other insurers, like Nationwide, are experimenting with AI in administrative roles.
Nationwide has introduced AI tools that summarize case details for claims representatives, allowing them to focus more on customer interactions. Unlike Allstate, however, Nationwide maintains that human agents are still better suited for delivering empathy and customer care.
Jeevanjee explained that Allstate’s customer service had previously suffered from excessive reliance on industry terms that often confused policyholders. For instance, representatives might reference “first notice of loss” or mention CCC Intelligent Solutions—Allstate’s repair cost estimation vendor—without offering sufficient explanation.
An example provided by Allstate illustrates the contrast between human- and AI-generated responses. A claims representative initially used the acronym “UPP inventory list,” while the AI version spelled it out as “Unscheduled Personal Property inventory list.” Such refinements help ensure customers fully understand their claims process.
AI’s ability to neutralize potentially tense exchanges is another key advantage. While human representatives may sometimes adopt an accusatory tone or question a customer’s account, AI-generated responses are designed to approach situations with a more understanding perspective. “Generative AI gives customers ‘the benefit of the doubt,’” Jeevanjee noted.
Additionally, AI-generated emails incorporate a more courteous tone. For example, AI-enhanced responses may begin with phrases such as “Thank you for your message. I understand your concern,” and conclude with, “If you have any other questions, feel free to ask.” By comparison, human-written messages often lack such reassuring language.
While Allstate insists this AI-driven transformation will not result in job losses, industry experts remain divided. A survey conducted by the International Customer Management Institute found that 35% of respondents believe AI will reduce hiring needs, while 38% think it will have little impact.
Beyond customer interactions, insurers are exploring AI’s potential for assessing risk, pricing policies, and evaluating claims. However, concerns persist about AI’s role in decision-making. United Healthcare recently faced criticism for allegedly denying claims based on AI assessments. It’s CEO was famously gunned down recently on a NY street with bullets that seemed to refer to the insurer’s refusal of claims.
Allstate, for now, is focusing on enhancing its AI capabilities. The company is evaluating whether AI can analyze images of damaged vehicles to determine the make, model, and extent of the damage.
Ultimately, the objective is to become a more customer-centric organization, Jeevanjee said. “If I think about the insurance industry in general, we haven’t really done a great job of being customer obsessed,” he admitted. “That’s really what I’m trying to drive.”
This shift aligns with broader trends in AI adoption across industries. Recent research from the University of Chicago Booth School of Business suggests that AI can surpass human analysts in certain complex tasks, such as predicting earnings changes.
The study found that GPT-4 outperformed human financial analysts in forecasting company earnings, achieving 60.35% accuracy compared to analysts’ 52.71%. These findings underscore AI’s potential not only in customer service but also in data-driven decision-making.
As AI’s role expands, companies will need to navigate ethical considerations and determine the best balance between technology and human expertise. While studies indicate that many customers still prefer interactions with real people, growing trust in AI suggests that preferences may evolve over time. The challenge for businesses, including insurers, will be leveraging AI’s strengths while preserving the human elements that foster trust and connection.
