Dive Brief:
Cigna unveiled a number of new digital tools on Thursday meant to improve customer experience with its health benefits portal, including a virtual assistant based on generative artificial intelligence.
The rollout — part of the insurer’s larger push to make it easier for members to access and afford the benefits they’re due — also includes a new tool to match patients to in-network providers.
Experts have raised concerns about rising adoption of AI in the healthcare sector due to the technology’s tendency to make mistakes. Cigna said its new features were developed with “rigorous” research and testing within an AI governance framework.
Dive Insight:
The vast majority of U.S. adults struggle with health literacy. Poor ability to obtain and understand health information contributes to worse health outcomes, more hospitalizations and a greater use of emergency care, according to research.
Low health literacy also costs the healthcare system tens of billions, if not hundreds of billions of dollars each year, analyses say.
Still, the U.S. has made little progress simplifying the complex layers of bureaucracy that make it harder to understand and receive care. Insurance is one issue. Many consumers are unclear what’s covered by their health plan, how much services will run them out of pocket and why they sometimes need to wait to get treatments approved.
That lack of understanding has contributed to rising frustration with insurers that came to a head late last year with the killing of UnitedHealthcare CEO Brian Thompson on the streets of Manhattan.
In the wake of Thompson’s death, major insurers pledged to roll back administrative barriers to care and become more accountable to their members. Cigna was no different, announcing in January plans to speed prior authorizations, expand access to patient advocates, simplify benefits and — later — to tie executive bonuses to customer satisfaction.
Now, Cigna is adding to its myCigna member portal, with at least one of the reforms leaning heavily on generative AI. The insurer’s new AI-based virtual assistant can answer common questions about health benefits coverage, claims and care options, according to Cigna.
More than 4 out of 5 customers found the AI assistant helpful in early testing, Cigna said in its Thursday release.
However, weaving AI into consumer communications might introduce new problems.
Generative AI models, which can create original text and images, are prone to errors, though the frequency of issues depends on the model. Models can cite incorrect sources, exhibit bias or make up information — an error known as hallucination. Hallucination rates, for example, range from 0.7% to 29.9% across a variety of different AIs, according to one analysis.
A Cigna spokesperson did not respond to questions about the model underpinning its AI chatbot by time of publication.
However, the AI can redirect customers to a human representative if additional help is needed, according to the insurer’s release.
Cigna also announced Thursday a new tool that gives customers a list of in-network providers based on their specific health needs, including different options for how care can be delivered.
Cigna plans to eventually integrate it into the virtual assistant.
The Connecticut-based payer also redesigned a cost tracking tool that breaks down deductibles, out-of-pocket expenses and bills for consumers, and new features to help members more easily upload claims and choose between employer-sponsored plans.
Many of the features are available today or will become available to all customers in 2026.
Like other stakeholders in the healthcare industry, insurers have been scrambling to integrate AI into their businesses to streamline operations and save money, whether through improving claims reviews, optimizing utilization management or other use cases.
However, providers and lawmakers have raised concerns about payers’ use of AI, pointing to incidences when insurers have used the technology to prevent patients from accessing medical care, allegedly in order to profit. Insurers have denied those claims.
