The intersection of intelligent automation and AI claims guidance was in the spotlight at the International Claim Association’s Annual Education Conference recently held in San Diego, California.
Flipping the traditional panel format, claims executives from Reliance Standard Life and Securian Financial put technology questions to business leaders from FastTrack, a leader in automation technologies for claims and underwriting, and EvolutionIQ in the panel discussion, “The Data Age: Drive Informed, Consistent Risk Management Outcomes By Capturing Key Disability Data with Intelligent Digital Automation.”
Of special interest to the audience was the way in which next generation artificial intelligence is now able to help claims teams take full advantage of the large amounts of complex data in each claim. A number of audience questions drove a free-wheeling discussion on how AI can understand overlapping medical codes and comorbidities, how it can read and gain insights from the conversational notes in a claims file, and how AI is able then leverage this to predict claim trajectories.
As EvolutionIQ’s COO and co-founder Michael Saltzman told attendees, “The machine learning in claims guidance software finds historical patterns and understands not only context, but cause and effect. Importantly, the system delivers an unmatched ability to read unstructured data – such as a medical professional’s extensive notes on a claim, an examiner’s notes taken while talking to experts or the claimant, legal notes, and other information that is in note form. As a result, it distills massive data sets in real time to understand recovery like a medical expert and predict outcomes with 95 percent accuracy. The AI partner then makes and explains action recommendations to frontline insurance examiners and adjusters so that they are guided to the right claim at the right time. Then it repeats the process daily to prioritize the highest-impact claims ready for immediate action.”
As Mike explained, even the most experienced examiners and adjusters have difficulty identifying claims ready for action due to complex case data and lack of real-time case views when new data hits the system. A key challenge is that the data contains both structured medical data (found in a properly labeled place in a core claims system or spreadsheet) as well as difficult to parse unstructured bodily injury and medical recovery data (the extensive, conversational medical and legal notes in each claim).
Being able to find, interpret and act on unstructured data is especially important because on average in each claim there are 2.5 medical codes that are in structured formats – but there are on average 14.5 codes relevant to the claim that lie hidden in unstructured, conversational notes formats that can run pages in length. This is what the natural language processing piece of the AI does.
Armed with these data-rich insights – and institutional knowledge from having examined tens of thousands of historical claim files – the AI is then able to predict claim trajectories with high accuracy. This in turn helps guide examiners and adjusters away from claims that are proceeding smoothly – and toward claims that could benefit from immediate actions.
“This is not a future state of technology,” Mike added. “Carriers and TPAs are using this technology now. It’s improving the claims handling process across the board, shrinking durations, and increasing claimant satisfaction.”