Automating Health Insurance Claims with AI

Artificial intelligence techniques can help companies lower costs, improve services and make better decisions. Chatbots for customer services and forecasting tools for better inventory management, machine learning can help the insurance sector improve claims management processes.

We asked founder of Caseblocks, Paul McGlynn for his insights on the introduction of artificial intelligence to claims management within the insurance sector.

Why are insurers deploying AI or machine learning techniques?

The main reason for a business to introduce machine learning techniques is to support and improve claims management by systematically identifying and correcting errors while avoiding ineffective interventions. It can also be used for detecting fraud, removing complex adjudication rules with a view to quicker decision making and a better claim experience for customers.

Let’s look at an example in the healthcare claims sector. For insurance companies checking health claims that come from hospitals are correct can be a paper-intensive, lengthy process involving hundreds of people depending on size of the insurer and the number of members it has.

Current processes typically involve a high number of claims being flagged by the health insurers’ rule book. Claims cases are checked manually by administrative staff. They examine claim data, policy data, benefit utilisation and the patient history if available, drawing on their experience to accept, reject or investigate.

Consider the valuable time and resources tied up in manual data checking that are being used within both the health insurers and also the health care providers.

Cognitive systems can help case managers to efficiently screen cases, evaluate them with greater precision, and make informed decisions.

How can machine learning optimise processes for insurers?

The machine learning techniques automate the process of  reliable identification of those, and only those, claims that are in fact incorrect.

AI approaches aim to identify only those claims for which the likelihood of successful intervention is high and, conversely, to route unobjectionable cases and those unlikely to result in successful intervention toward fully automated background processing so that administrative staff can effectively focus their capacity on cases that require review.

How does Caseblocks deploy machine learning in the claims management process?

We support insurers by providing a platform with the capability to automate the entire workflow. Our approach to technology comes from a deep understanding of processes, information flow and the need to manage both structured and unstructured data. Every Caseblocks deployment starts with a discovery phase, a pilot to develop a proof of concept, then testing of the derived machine learning benefits with further deployments once successful.

What are the challenges for insurers considering and evaluating machine learning for their organisation?

Good question. We know that insurers have been using highly sophisticated data analysis techniques for years, so in principle using machine learning to optimise their processes is not a huge leap. On the other hand, a platform to manage all the data they require to take advantage of machine learning techniques is more complex.

Gartner estimates that most insurance companies process only 10 to 15% of their structured data that is sitting dormant in traditional databases.  That’s before we start to look at the unstructured data such as photocopied medical notes or lab test results, which can provide valuable insights.

The question then becomes how do insurers not only unlock the value of their structured data but also gain valuable insights by bringing their unstructured data into the same platform.

The transformational element of a AI project is the platform that interfaces with structured and unstructured data.

What skills will insurers need to bring into machine learning projects?

It’s very useful to get people with a data science interest into the start of a deployment of any data platform. We recently went through a discovery phase with a health insurer and working with the data scientists from the outset is key.

Analysing this unstructured data and using it to drive better business decisions requires advanced data science techniques. Emerging data analytics technologies centred on machine learning bring order and purpose to this unstructured data so that it can be more effectively mined for business insights.

One major benefit of machine learning is that it can be effectively applied across structured, semistructured or unstructured datasets. It can be used right across the value chain to understand risk, claims and customer behaviour, with higher predictive accuracy.

What are the key opportunities for insurers?

The potential applications of machine learning in insurance are numerous: from understanding risk appetite and premium leakage, to expense management, subrogation, litigation and fraud identification.

An advantage for insurers with strong data backgrounds will be the ability to bring in their own underlying data assets to merge with the capabilities of these new predictive models – generating unique insights and opportunities.

Tips for deployment of AI

It’s critical to consider the entire workflow when deploying a new system for machine learning. The conditions we at Caseblocks help to put in place for our clients aim to ensure that the system works reliably within the existing day to day processes of the organisation and that the ‘transactional’ workload of employees is indeed reduced as planned so they can focus on specific cases or more meaningful tasks. Technology is the foundation for the data, but it does not sit on its own, at the very beginning of any project you need to ensure that the team are ready for both a digital and cultural transformation.