I had an opportunity to talk to Joanne Wright of IBM not long ago. Ms. Wright is a Vice President of Supply Chain; she has a charming Scottish brogue. Ms. Wright has overall global leadership of the supply chain team across procurement, manufacturing, demand management, quality, and supply chain risk management for mainframes, POWER-based servers, storage products, and client solutions across the company’s cognitive and cloud platform product lines. IBM manages more than $30 billion a year as part of its supply chain, with thousands of supply chain managers and analysts worldwide.
Joanne Wright, VP of Supply Chain at IBM
IBM is using Watson and huge troves of weather and location data – IBM acquired The Weather Company last year – to improve their supply chain risk management program. IBM has regional control towers with a small, core centralized team in New York with a global control tower. IBM’s regional teams link to the Geodis visibility information to do their own planning. IBM plans and executes shipments to 170 countries for IBM’s clients from across their supply chain network of IBM manufacturing, original equipment manufacturers (OEMs), and original design manufacturers (ODMs). They use regional control towers and a small, core centralized global team in the US to track shipments across the global network.
After IBM acquired the Weather Company its’ goal was to use cognitive technologies to create sophisticated models that analyze huge troves of weather and location data. By combining weather data with advances in atmospheric and computational sciences, these systems produce more reliable weather forecasts, including the location-specific impacts of storms, hurricanes and typhoons. IBM is using these weather forecasts to allow supply chains to more quickly anticipate and deal with extreme weather, which can wreak havoc on any supply chain.
The AI is combined with streaming data from social feeds, news reports and weather data from the Weather Company while Watson APIs extract relevant information from social media and news feeds into their control towers. Social media is used by creating a large set of pertinent terms – like “hurricane, “port strikes”, etc. – that can impact supply operations. These terms are continuously searched for by a resiliency engine and when detected sent to the relevant managers after they are validated as being impactful. IBM is continuing to analyze these social feeds to eliminate the noise of “false positive” alerts by determining which of the feeds have the best predictive value. You can view this video if interested in learning more details on the Watson weather technology.
Their cognitive engines also compares a storm against historical data. Watson forecasts which typhoons or hurricanes are more likely to hit an IBM location or one of their key suppliers, and which ones won’t. If there’s a risk, IBM can purchase components ahead of time and move them to a safer location, evacuate employees at a site until a storm passes, or look for alternative sources of a supplier will be down for an extended period.
One example of how IBM has used these capabilities occurred during last year’s Hurricane Patricia, the second most-intense tropical cyclone on record. This cyclone picked up steam so quickly that it left much of Mexico’s coastline unprepared. But 150 miles inland, IBM’s Guadalajara production center was already alerted by the cognitive supply chain system and was ready for the storm. As expected, the storm struck north of the city, causing minimal damage to production operations, although IBM evacuated the center as a precaution.
It is worth noting, however, that despite advances in forecasting, IBM did evacuate the plant even though their forecast correctly said the cyclone would track far enough North of the plant to avoid damage. A weather forecast is still a forecast; and the risk of making the wrong call here were too high not to have ordered an evacuation. But, the solution allowed IBM to make contingency planning decisions earlier and faster than they would have otherwise. For example, some inbound shipments were routed to the US and then shipped back to Mexico after the hurricane passed.
Aside from catastrophic events, Watson can also detect weather that can impact transportation including wind speeds, rainfall. IBM is also using traffic data.
The journey toward using weather data is part of a transformation IBM is undergoing to capitalize on cognitive computing. Cognitive computing is IBM’s flavor of artificial intelligence (AI). This was the technology the Watson supercomputer used to win a Jeopardy tournament in 2011. Today, Watson is a cloud service that can analyze mountains of data. IBM is also using these capabilities on the service side of their business to help their customers do everything from finding better cancer and diabetes treatments, to helping spot retail trends so stores can anticipate demand. When it comes to weather analysis for supply chain resiliency, they call the service “Supply Chain Risk Insights powered by Watson.”