How machine learning is transforming insurance claims
Here are some of the use cases for machine learning in claims as well as the challenges limiting adoption.
Machine learning (“ML”) has been one of the most prolific areas when it comes to high-impact use cases for the insurance industry. And within insurance, claims management offers one of the most promising areas to apply this technology due to the large amount of data available to train algorithms and the consistency of principles applied in the claims assessment process. Here, we look at some of the use cases for ML in claims and challenges limiting adoption.
Focus on fraud
Fraud detection is perhaps the area where we see the most advanced adoption of ML among insurance companies. Start-ups such as Shift Technology, Friss and Owl Labs have seen strong demand from carriers and attracted significant capital from investors to support their growth. These tools function by applying cutting-edge data science to large historical claims data sets, enriched with third-party data.
Insurers have been quick to adopt ML-based fraud detection strategies because they can often deliver the most immediate and tangible return on investment. Many fraud teams within claims rely on rules-based approaches (often inside an adjuster’s head) and tend to focus on those claims where the scope for fraud is most obvious, potentially neglecting larger, more complex fraud cases. In contrast, ML can often detect more subtle patterns of fraud that might not be visible on an individual claim. For example, is a single motor body shop regularly overcharging across multiple claims? Is a physician’s office regularly diagnosing patients with whiplash in low-impact accidents? Patterns like these may only become apparent when reviewing all of the historical claims associated with an individual policyholder or vendor.
Automating the assessor
An additional area of significant innovative activity has been in damage assessment. The development of computer vision to analyze photographs of claims has been the central innovation of multiple startup companies, particularly in personal lines.
In motor claims, companies like Tractable and Snapsheet allow policyholders to submit photos of their damaged vehicles and generate a claim estimate without a human assessor viewing the image. Their models are trained on historical data for similar vehicles and are often accurate enough to settle a claim instantly or make an informed decision on how best to take a claim forward. Computer vision assessments speed up decision-making, reducing claims leakage (paying too much to settle a claim), and improving the customer experience. Expert assessors still have a role but are directed to the more complex cases where there is a low degree of confidence.
We are seeing similar products emerging in property damage assessments from start-ups like Flyreel (acquired by LexisNexus), Hover, and Hosta Labs.
Document processing
A more nascent area that we believe has enormous potential is the use of ML to automate the processing of complex documents. Claims professionals today are buried in documents. Even simple claims can feature damage reports, doctor’s notes, multiple invoices, emails, and texts — all of which contain crucial information in unstructured or semi-structured formats.
Natural language processing (NLP) and computer vision technology have the potential to reduce manual data input significantly. ML applications can look at an invoice, for example, and extract individual line items, payment information, and invoice numbers. At the end of the process, an accounts payable professional only has to approve the invoice with all the necessary context at their fingertips. Smaller claims that meet certain criteria and with no “red flags” can be approved without any human interaction.
Start-up companies often train their models on a single class of documents (at least initially). Hypatos is a leader in invoice processing. DigitalOwl focuses on reviewing medical records. Groundspeed Analytics started as a loss-run specialist. Insurers can put together these tools to extract all necessary information from documents so claims professionals can focus on decision-making, not data input.
Barriers to adoption
In short, we see new companies forming that are using ML to bring efficiency to claims management. Perhaps the single biggest challenge these companies face today in adoption by incumbent insurers is integration into existing systems.
Data is the lifeblood of any machine learning model. ML applications need historical data to train and tune models and once they are up and running, they need rapid access to new claims data to be effective. At present only a small number of carriers have the technology to deploy ML models in their claims operations.
There are dozens of potential ML use cases that can bring efficiency to the claims management process. However, many start-ups get so bogged down in integrations with carriers in connection with initial pilots that the projects never get beyond the POC stage. To fully unlock the potential of ML in claims, insurers will need to transform core IT systems.
Based in London, Jack Prescott is a senior associate at MTech Capital, a venture capital fund focused on the insurtech space.
See also: