The following comment made by a speaker at the annual Accord Insurance Conference in the US sums up the mood in the Insurance industry:

"The next few years will be an interesting period," summarized Chung, "Money is tight, competition is fierce, and technology is moving."(Accord Conference 2003 Report)

So on one hand there are regulatory and standardization pressures on the industry and on the other, optimizing performance and streamlining internal processes is essential because of growing competition and increased risks due to disasters like 9/11 and the SARS epidemic. In the these troubled times it is important that the industry moves to a new level of technology adoption. Insurers should look out for technology to increase efficiency and control costs.

One area where insurers could control cost is the area of claims handling and management. One short-term way out of this situation is to outsource the claims processing to countries where labor-cost are really cheap - as seen in the recent upsurge in the BPO market. Adoption of technology in this area is slow, in January 2003, A.T. Kearney, one of the leading technology research firms in BPO segment, reported that insurance underwriters could save up to 3 billion dollars by shirting to more intelligent claim processing technology. A brief online survey of the technologies employed by the insurers suggests that the present claim-processing is done either manually or with database-driven rule-based technologies (e.g., IBM's UBClaims, ScanSoft and VisionNet).K-Praxis thinks that the new level of technology automation in the insurance industry could come from the ability to intelligently process data by using various AI, machine learning and statistical/semantic NLP based technologies.

K-Praxis reviewed some of the more advanced machine learning based claim processing technologies. Here is what we found:

MindBox Intelligent Text Analyzer

Mindbox's technology automates the claim processing by extracting and "reading" relevant information from the forms(not only from the numerical and logical fields, but also from text-rich free form text fields),identifies actionable information important for decision-making. Mindbox uses two different mythologies (apart from normal rules and pattern matching techniques used by claim processing software): tagging, itemizing and interpreting sentences using text analysis; and Case Base Reasoning (CBR) technique. CBR, is an advanced machine learning technique which can automate and assist in complex decision making procedures.

Fair Isaac Claims Advisor For Subrogation and Capstone Decision Manager For Claims

Fair Issac's technology fuses analysis of structured, semi-structured and structured data through advanced text analysis, neural network predictive models that use common data elements such as body part, nature of injury, accident cause and accident description text to process, to interpret claims in order to aid the decision making potential of an insurer.

Megaputer PolyAnalyst

Megaputer, a very well known data and text mining software vendor has forayed into insurance claim management market with introduction of semantic-NLP based PolyAnalyst Text Analysis. PolyAnalyst extracts relevant text data, categorizes this text data in relation with Megaputer's Link Analysis and Decision Forest - advanced data mining capabilities embedded in its PolyAnalyst 4.5.

Solutions enlisted above are a representative sample of a growing market for software and technology to tackle the vast sea of unstructured information available through insurance forms and claims.

Important Note: K-Praxis' Technology/Product Review section focuses on technologies and vendors that are making a significant impact in the field of intelligent and automated information management. These presentations do not constitute a product endorsement by K-Praxis