A.I. and Computer Vision Can Help to Remotely Monitor Operations

24X7 Alerts & Analytics Flag Problem Areas Without the Need to be Onsite  

The Coronavirus has made major production disruptions in China with over 80,000 reported cases and 3000 deaths. But factories are also being impacted in Japan and South Korea.  Together the three East Asian countries represent 38% of global manufacturing. 

Source: https://gisanddata.maps.arcgis.com/apps/opsdashboard/index.html#/bda7594740fd40299423467b48e9ecf6

In the last few weeks, major outbreaks have been reported in most of Europe which will lead to further manufacturing and distribution disruptions.  Disruptions will also expand in North America.

Source:  https://time.com/5800901/coronavirus-map/

Today’s global supply chains are much more fragile than in the past. When I began in supply chain in the 1970s, there was a preference to build in safety stocks into inventory planning and source locally whenever possible.  With the advent of Lean based on the Toyota Production System, the focus changed significantly to a just-in-time philosophy. The goal was to bring parts into the factory at the last possible hour and day. This impacted warehousing, transportation, and third-party logistics as well. Running parallel with the growth of Lean is the obsession to chase the lowest global labor costs – a practice known as labor arbitrage.  Most often it meant moving to suppliers in Mexico, East Asia, and Southeast Asia.  

The Coronavirus (technically designated as COVID-19) has created a shock to global supply chains. The first reaction of supply chain professionals would typically be to jump on the first flight available to get on site at their suppliers. But travel of all kinds is now considered a high-risk proposition. 

Even without a pandemic supply chain shock, site visits to distant and offshore suppliers have major limitations. Suppliers are well versed in how to put on the best possible face to their customers’ sourcing teams. The problem is obvious, once the visitors depart, the suppliers revert to their normal operations, which may not be at the level needed to support their customers. Even monitoring local operations can be challenging.  It is not possible for a manager to be everywhere at once, this is compounded for those responsible for very large and dispersed operations – major distribution centers can run over one million square feet, large surface mining and quarry operations can cover a square mile. 

A much more effective remedy is to use artificial intelligence and computer vision to create a 24×7 monitoring and alerting system. Unlike an occasional and planned site visit, computer vision never stops watching and analyzing areas of interest. Unlike a passive/dumb camera system, there is no need to go back and review hours of video coverage.  Smart cameras use algorithms to send alerts whenever issues occur.

An example of using computer vision is the monitoring of a supplier to verify that a production line is actually running for an entire shift and that a set number of workers are active in the production area.  Alerts can be created to flag when the line stops for a given number of minutes and when workers are missing from planned staffing levels.  

Another example would be to monitor for process violations, such as skipping a testing station. When a violation occurs, an alert would be created in near real time.  Experience has shown that such monitoring will almost always result in improved process adherence and processing time efficiency. 

Examples in distribution would include the time to load and unload trucks, the wait times to enter yards and docks, and the inventory levels in stockrooms. 

There is a continuing trend to move manufacturing from China to Southeast Asia, Mexico, and back to the United States. There is also an increasing likelihood that Corona-like epidemics and pandemics will no longer be Black Swans. Both of these trends make it very timely to deploy an A.I.-based computer vision monitoring and alerting system. This makes good sense whether your operations are down the street or thousands of miles away. 



Anthony Tarantino, PhD

Adjunct Professor, Santa Clara University – Operations and Finance

Six Sigma Master Black Belt, Certified Scrum Master, CPIM (APICS), CPM (ISM)

Senior Advisor to Atollogy