Aflac takes on claims challenges to scale AI efforts

For Aflac, which offers supplemental insurance coverage to greater than 50 million individuals worldwide (and is well-known for its duck mascot), delivering AI at scale throughout the group has grow to be a high precedence because the pandemic.

Aflac has been compelled to speed up its digital transformation, together with synthetic intelligence (AI), because the pandemic severely challenged the corporate’s conventional in-person, impartial agent/franchise enterprise. The trick, nevertheless, has been selecting the perfect AI use circumstances amongst competing priorities, says Shelia Anderson, who joined Aflac as CIO final July. 

“We’re enthusiastic about the enterprise problem and outcomes we’re on the lookout for,” she instructed VentureBeat. 

In relation to AI and machine studying, that features specializing in the general viability and desirability of the corporate’s fashions, and asking questions equivalent to: Is the mannequin wanted by the enterprise? Is it fixing a particular enterprise want? Does the corporate have the technical options it wants? How lengthy will it take for the mannequin to convey worth to the enterprise? 

A transparent alternative to automate claims

Aflac has lengthy had a deal with massive knowledge. Immediately, Aflac helps brokers and brokers with AI and ML fashions that assist in suggestive promoting, flagging at-risk accounts and figuring out dormant accounts which might be candidates for reactivation. 

However growing an answer that would scale AI throughout the group has been a excessive precedence since 2020, stated Anderson. Simply final yr, the corporate rolled out what Anderson calls its first vital AI-driven platform that makes use of AI and ML to rework how Aflac processes claims.

The platform consists of a set of fashions educated on enterprise guidelines tailor-made to the corporate’s numerous product traces. The objective is to automate routine processes, permitting the corporate to pay claims extra shortly.

There are three essential parts to the platform:

1. An AI-based doc digitization pipeline to routinely extract, classify, annotate and index proof-of-loss paperwork
2. Information graphs to map extracted data from paperwork for a greater context of processed knowledge.
3. An end-to-end, AI-based claims processing workflow for adjudication throughout totally different traces of enterprise, permitting for absolutely automated or assisted, error-free, human-in-the-loop processing.

“This helps our clients to be adjudicated sooner and with extra accuracy,” Anderson stated, stating that earlier than the AI answer was applied, about 46% of Aflac claims weren’t absolutely automated. 

Aflac has many alternative declare sorts, she defined, however one of many first clear alternatives to scale AI was across the firm’s wellness advantages. These are included in most of its accident, hospital indemnity and most cancers insurance coverage insurance policies. Primarily, Aflac pays clients cash for getting yearly checkups and medical screenings equivalent to physicals, dental exams and eye exams. 

It turned on the market was a excessive quantity of lower-dollar payout claims requiring time-consuming buyer interactions. 

“For easy claims that don’t require proof of loss, like wellness claims, we need to pay out shortly,” stated Anderson. This “permits our buyer care specialists to care for our policyholders [who have] extra advanced conditions.”

Scaling the AI platform

Now, Aflac is working to scale its claims automation platform to different varieties of claims. 

“The advantages that the enterprise case has confirmed are improved buyer ease, decreasing our ache factors by means of the journey, and growing our touchless claims, which was a profit to our inner workforce in addition to our claimants,” Anderson stated. “Streamlining with a rules-based AI reduces error charges and frees up our sources to allow them to deal with extra vital claims the place individuals may very well want to listen to a voice on the opposite finish of the telephone, possibly coping with extra extreme health-related points the place that private contact is required.” 

Anderson stated she believes Aflac has solely simply hit the “tip of the iceberg” in relation to implementing the platform. She has plans to broaden the identical functionality throughout the group in 2023. That, she identified, is the worth of getting a mannequin that works effectively, one which solves a primary problem and takes benefit of a chance within the market.

“You’ll be able to take that and stamp it throughout your different traces of enterprise with the same downside,” she stated. “So we’re taking this and increasing it in our accident and hospital traces of enterprise, and we’re additionally including different capabilities sooner or later round most cancers, dental and imaginative and prescient.”

As well as, she added, there is a chance to increase these AI capabilities past the claims course of, to any use case that must be automated primarily based on prediction.

Aflac’s greatest AI scaling challenges

Apart from prioritization, one of many greatest challenges in scaling any AI effort throughout Aflac is getting participation from numerous organizational entities, Anderson stated.

“For instance, our accomplice that runs the analytics aspect of our enterprise has a front-end staff,” she defined. “We’ve a back-end knowledge staff after which we’ve got enterprise groups that we work with as effectively. So managing and prioritizing throughout that ecosystem, whether or not it’s AI or whether or not it’s one other enterprise initiative, that’s all the time going to be one thing that may be a problem for us.”

As well as, in a high-demand area like AI and machine studying, attracting and retaining expertise with the suitable ability set is a significant problem. “It’s one thing all of us have to remain laser-focused on,” she stated. 

Making use of AI to enhancing buyer retention 

General, Aflac’s claims automation platform has helped with customer support and buyer retention, Anderson stated. 

It’s about “how we spend the time that we’d like for these highest-priority clients and claims whereas automating others,” she stated. “I feel that customer support goes to be key in leveraging AI sooner or later.” 

That stated, she added that she believes permitting some AI capabilities to mature has been an vital a part of Aflac’s journey — taking time to ensure it doesn’t take useless dangers with buyer interactions. 

“If you wish to be first to market with one thing, after all, that’s only a threat you’re going to need to take,” she stated. “However for Aflac, I imagine that permitting a few of these capabilities to mature was positively a part of the journey.”