the science of figuring out whats missing

These days, companies are collecting huge volumes of data about customers, products, employees, and services. One of the big challenges is storing and trackiing it in superior ways so they can do business effectively and comply with regulations. Our IBM Research scientists in Japan, New York, Zurich and elsewhere are working hard to produce major advances in this area, which we call managing business integrity.

An important piece of the puzzle is just now coming out of our Tokyo lab. Working on behalf of Japanese insurance giant Dai-ichi Life, IBM researchers have produced a new semantic analysis tool that the insurance company expects will save it millions of dollars per year by improving its ability to diagnose and fix problems that emerge in its business processes, such as processing claims. It’s a so-called First-of-a-Kind project where researchers accelerate the development of technology by working closely with a client on its real-world problems.

The insurance company converted from paper-based claims processing to a digital system, but couldn’t gather sufficiently detailed information about the actions taken at each stage of the process. When each employee involved in the process completes their discrete task and passes the claim on to the next person in the workflow, the actions they take are recorded as event logs. But the workflow system does not record complete information in the logs, making it harder to resolve quality and efficiency issues. “Invisibility was the problem,” says Satoshi Hada, one of the lead researchers on the project.

IBM scientists developed a new software technology that uses semantic analysis to make it easier for Dai-ichi Life to get a clear view of what’s was going on inside its claims processes. By examining event logs from earlier and later points in the processes, the system can figure out what action was taken at a particular step, making it possible to spot the root causes of problems and fix them.

The project was started just over a year ago and was completed in March. Now researchers are continuing to develop the semantic analysis technology so it can become an element of IBM’s Insurance Operations of the Future initiative–a complete data management architecture for the insurance industry. The technology has a lot of potential beyond the insurance industry, as well. Wherever data is missing, people will need help finding it.