7 AI Use Cases in Health Insurance

According to a recent Gartner survey, leading organizations expect to double the number of artificial intelligence (AI) projects within the next year and more than 40% expect to deploy AI solutions by the end of the year. For health insurers, AI offers an opportunity to streamline processes, improve fraud detection, and deliver a better customer experience. In this post, we’ll examine some AI use cases in health insurance that can help achieve organizational goals.

What is AI?

Let’s create a working definition of AI. Essentially, it’s any computing technology that exhibits some behaviors associated with human intelligence, like planning, learning, reasoning, problem solving, perception, and creativity. Several subcategories can be described as AI, like machine learning, deep learning, neural networks, and natural language processing.

At its core, AI takes common human tasks that require some type of decision-making and leverages computing power to complete the task.

Implementing AI

If you don’t have an AI strategy document, that’s the first step in your AI journey. Key features of a strategy document include:

  • Define AI – Just like we did earlier, define what AI means so everyone in the organization has the same working definition.
  • Analyze the potential impact of AI – Include an analysis of the potential impact of AI on your business at a high level. If you can connect those impacts to your business goals and objectives, do so. AI isn’t always about hard ROI; you can also align with improvements in customer service scores or even with sales objectives.
  • Keep a running list of prioritized AI test cases – Your strategy document should contain a list of potential AI projects — including all phases of design, testing, and deployment — you can use as a reference and to get buy-in from stakeholders and decision-makers. Include information about the issue the AI solution solves, the estimated expenditure, the expected return on investment, and any other factors that would be relevant to the AI use case.
  • Define your prioritization process – Prioritizing your AI projects isn’t easy. I’d start by creating a list of criteria you’ll use to prioritize potential AI use cases. In a lot of cases, balancing man-hours to test and release the AI project with the return on investment will help you understand which projects require your focus. If you’re just starting, look for test cases that offer immediate returns — however small — with minimal effort. Starting small and being successful will help you gather steam for larger projects with more significant ROI.
  • Define how you transition AI tests to production – What constitutes a successful test? How do you transition a test to production? For many insurers, testing an AI solution is the easy part. Transitioning that solution to a production environment and ensuring it operates as designed is difficult. So spend some time mapping how to take successful tests live.

Adding a strategy document will help you formalize your process, inform stakeholders, and give you guiding principles that will help keep AI central to your business.

7 AI Use Cases in Health Insurance

Now that you’ve defined AI and formulated a strategy, let’s examine seven potential AI use cases in health insurance:

Claims Processing Automation

In this McKinsey article about claims processing, they indicate that their experience shows that one in ten claims is incorrect and the claim can be challenged by the health insurer. But finding those claims is difficult. Up to 70% of claims are flagged as unusual and then passed to administrative staff to review in detail.

Adding AI to this process can help limit the initial number of claims flagged as unusual. As a result, fewer claims get sent to administrative staff to review. Which is an important point: AI usually isn’t a replacement for humans. It augments their ability to do more where their intelligence can outperform a machine or machine learning. In this case, creating AI that can take a set of rules, see how they’re adjudicated, and then take in a growing amount of data to learn which are most likely to need a human review can significantly reduce the amount of time staff spends reviewing claims.

Prior authorization

For providers, dealing with multiple insurance companies and their differing guidelines for prior authorization can be an administrative headache. For insurers, that often means incomplete data and wasted administrative time tracking down the data required to decide.

Infinx helps providers and insurers fix the prior authorization mess by automating their prior authorization workflows. The solution enables providers to electronically submit the request to insurers. It leverages AI to determine which procedures require authorizations, the information required for those authorizations and then leverages AI for follow-up. It’s a better experience for insurers who aren’t inundated with manual submissions likely missing required information, streamlines provider administration, and improves the patient experience because procedures can be scheduled faster.


Today, most insurers are using chatbots in some form to engage members. However, many are likely leveraging standard chatbots that are programmed to perform very specific functions. But AI opens up a world of possibilities, from more tailored insurance shopping experiences to better customer engagement.

Leveraging buying behaviors and satisfaction data, you can create data-rich chatbots that recommend the best plans. Those recommendations can be based on an individual’s needs as well as the real-life experiences of similar members. From a customer service perspective, building AI-powered chatbots that can learn and process the subtleties of languages over time to improve understanding of context can create better engagement. Many insurers today leverage chatbots to determine where to route inquiries. AI offers an opportunity to offload a vast majority of those conversations entirely, saving millions in administrative costs.

Call routing

We’ve all been there: We wait on hold for 15 minutes after answering a series of prompts. When we finally reach a human, they ask that same series of prompts again. Then transfer us to another representative because we were placed in the wrong queue by an outdated IVR system.

AI can significantly enhance the call routing process. In addition to department-level routing, AI can leverage data to route calls to those with specific skills within a department. AI can leverage the context of the conversation to perform more intelligent routing so the right agent gets the right member.

The future of AI in call routing involves personality matching. Each of your representatives likely has a different communication style and skill set that appeals to different subsets of members. Some are poor with negative callers; some are excellent at managing angry members. Intelligent AI can identify those angry callers. Then, using AI, match them with the appropriate agent, likely increasing customer satisfaction, and improving call resolution rates.

Fraud/abuse prevention

Highmark employs a Financial Investigations and Provider Review (FIPR) team to detect fraud and abuse. That team, consisting of registered nurses, investigators, accountants, former law enforcement agents, clinical coders, and programmers, performs audits to detect unusual claims, coding reviews, and investigations to detect fraud and abuse in payments to providers.

But they also leverage AI to identify abuse. Unlike their human counterparts, AI can review a vast amount of data. During that review, it can identify changing behaviors sooner than traditional methods, which rely on established rules. The team generated more than $260 million in savings associated with fraud, waste, and abuse in 2019.

Leveraging AI to identify fraud and abuse not only saves money on the front end but can become a key selling point for an insurer’s sales team. Those advanced techniques can become a real competitive advantage that lowers member costs.


Certifi’s health insurance premium billing and payment solution leverages AI to streamline operational processes – like payments sent to lockboxes without accompanying account information. Machine learning can help apply those payments to accounts, eliminating a manual review. It’s another simple way to leverage AI to streamline operational processes.

Early Intervention

Cigna’s corporate venture fund has invested in GNS Healthcare, a company with an AI platform that transforms data into personalized treatments, and Prognos Health, a company that provides health monitoring AI to track and predict disease. Prognos can leverage data and AI to help Cigna predict when health events may occur for members, Then, they intervene to provide early care, potentially improving health outcomes, and reducing costs.

Realistically, this approach is the future of healthcare. Today, healthcare is primarily reactive — members get sick and seek care. The future of healthcare is much more proactive, leveraging data, AI, and predictive models to understand individual patients and personalize intervention and healthcare. For insurers, that model likely leads to better patient health outcomes and reduced cost. As the old Benjamin Franklin saying goes, “An ounce of prevention is worth a pound of a cure.”

Certifi’s health insurance premium billing and payment solutions help healthcare payers improve member satisfaction while reducing administrative costs.



Emerging Technology: A Health Insurance CIO's Guide

Related Posts

Start typing and press Enter to search

Get New Posts in Your Inbox!

Skip to content