Predictive analytics has become a driving force in the use ofdata by insurance carriers, but many issues remain, including theproblem of sorting through unstructured data to improvedecision-making.

Participants in the virtual conference "Facing the Future ofInsurance," sponsored by Tech Decisions and its sister publicationsClaims and National Underwriter P/C, were treated to a discussionof the issue by Kaleb Adams, assistant vice president, claimsresearch, for The Hartford, and Karen Pauli, research director inthe insurance practice at TowerGroup.

Those who missed the initial presentation still have anopportunity to hear Adams and Pauli by going to the conference Website and registering at this Web page:http://vts.inxpo.com/scripts/InXpo.nxp?LASCmd=AI:4;F:QS!10100&ShowKey=1278

Adams explained traditional data alone does not paint a fullpicture, particularly in the investigation of fraud. Among theevidence that will help a carrier settle a claim are: evidenceuncovered by the claims handler's investigation; witness statementsand employer reports; and medical reports from doctors.

Pauli pointed out the need for an enterprise data strategy thatgoes beyond traditional structured data (data warehouses, datamarts) and hybrid data (comment field, form fields) to includeunstructured data such as PDF documents, e-mail notes, Internetmessages, voice-mail messages, and digital photos.

In addition to this session on predictive analytics, there isanother technology session available in the virtual conference on"Speed and Ease: Underwriting at Its Best." Speakers in thatsession include Robert Asensio, senior vice president and CIO atDarwin Professional Underwriters, and Donald Light, senior analystin the insurance practice at Celent.