When Watson Meets Procurement – IBM Uses Big Data to Tackle Big Supply Challenges


One of the presentations at Coupa Inspire this week that I was particularly inspired by did not come from Coupa at all, but from Michael O'Leary, director of procurement offerings and strategy at IBM Commerce. Michael is a former procurement and supply chain executive within IBM. If you are interested in the Cliff Notes version of IBM’s procurement transformation, read the article from yesterday, IBM and Procurement Transformation: By the Numbers, Risk Management and More.

During the second part of his talk after the IBM transformation story, Michael shared some of the initiatives IBM is working on around big data and procurement, leveraging Watson Data Explorer – IBM’s big data solution – alongside procurement and related supplier data.

Risk Management, Meet Watson

The first example Michael cited was a scenario in which Watson helps parse through unstructured data to drive a new level of supply risk management intelligence by taking “soft facts” and information from Twitter, news feeds, social networks and other sources and lining these up against more traditional risk management data feeds (e.g., D&B, systems data, etc.).

Under this approach, Watson aggregates these soft facts up to the KPI level and then layers this information into the IBM Emptoris Supplier Lifecycle Management (SLM) solution. The result: a nearly n-tier reach into all of the facts or clues that may have an influence on supply risk scoring with a given supplier or supply chain outside of what has traditionally been used in the past to gauge performance.

Ask Watson – No Question Too Big or Little

The next Watson initiative that Michael shared was the use case of “Asking Watson About Procurement.” The concept behind this scenario is that organizations have reams of data that is buried in contracts and transactional systems (internally) and externally related information (e.g., regulations) that all tie back to information that is usually not leveraged in procurement activities, at least not everyday.

The idea with Watson in this regard is that the solution can leverage its natural language processing capability to answer questions based on free text entry (regardless of how simple or complicated the question or answers are). The ability to query off of natural language could enable, for example, the ability to ask Watson about a question such as: Does [my organization or another] have greater diversity spend than others compared with peers?

Build My Briefing Book, Watson

Another example that Michael provided might best be described as the category/supplier “briefing book” on steroids. The idea behind “the complete procurement overview” is to aggregate information on a real-time basis when anyone within procurement (or outside of it) needs a briefing document or related auto-generated deliverable.

The real life use case here might be an executive or category manager going into a performance review with a supplier – or the same individual wanting more information when talking on the phone with a supplier to provide in context, on demand.

No doubt, IBM is up to something with Watson, and as IBM deploys these capabilities in customer environments, you can be sure that we’ll be among the first to cover it here.

Discuss this:

Your email address will not be published. Required fields are marked *