In his three part series, ServicesAngle editor Bert Latamore examines the transitions that IBM’s famous AI platform, Watson, underwent in the past two years. The synthetic Jeopardy!  champion transformed from a bulky supercomputer into an artificial intelligence system that leverages powerful algorithms to diagnosis patients and save lives.

Medical information is complicated, unstructured, and aplenty: exabytes upon exabytes of genetic data, lab results, patient records and even handwritten doctor’s notes. In other words, big data.

A couple years ago, IBM’s Watson team began collaborating with experts from the Memorial Sloan-Kettering and a company called WellPoint to apply the system’s natural language engine big data analytics, specifically diagnosis of difficult to treat cancers. Why cancer? In the words of IBM Fellow Rob High, “we could make a difference.”

In its present firm, Watson is considerably smaller and 240 times faster than the original machine. It also consumes less power, and comes equipped with sophisticated software that piles connect the dots between medical data as effectively as a human doctor. High says that this software, which emulates the way the human brain contextualizes information, is the heart of Watson.

IBM’s AI researchers spent two years talking with cancer experts in order to develop the context patterns that are needed in this field. Big Blue even had oncologists compare Watson’s findings with their own to see if the underlying algorithms are accurate enough for production environments.

When Watson was finally mature enough for used in the field, IBM transformed this technical marvel to an enterprise-ready product. The company is offering the platform as an SaaS solution that is hosted on its centers, and it does not retain clients’ confidential information in order to protect patients’ privacy and ensure compliance with industry regulations.