Agility. Adapting to changes quickly and pushing ahead. During the coronavirus pandemic, businesses have had to react almost overnight to unpredictable changes out of their control. To survive, they’ve had to keep up with the changes and adapt to them, often with machine learning and AI tools.
While almost all types of businesses have been hurt by the coronavirus, those with industrial supply chains have been among those hit the hardest as manufacturers shut down plants and retailers close.
AI for supply chain predictions can be useful to help enterprises adapt quickly to ever-changing circumstances and enable them to better weather the pandemic.
Bringing AI to the supply chain
Machine learning and AI for supply chain “allows you to very quickly adjust to new data” said Bill Waid, general manager of decision management at FICO, the credit score company.
Using machine learning, enterprises can make predictions during an unpredictable time, including making demand forecasting models and logistics models based on small amounts of actual data or hypothetical data. To do that, humans need to work directly with machine learning models, inputting data and predictions proactively and picking from hundreds or thousands of potential futures the models predict, Waid said.
Creating hypothetical scenarios, such as a situation in which a factory is closed or only a fraction of a retailer’s products sell, enables businesses to plan for the worst or most likely. Adding in even small amounts of data, such as demand data captured over two weeks of social distancing during the coronavirus, businesses can add some predictability to their forecasting.
For example, a model, based on the last two weeks, could predict that demand for toilet paper will skyrocket by 40%, while the demand for new clothes will plummet by about the same amount, enabling a business to adjust its production and shipping routes.
The business could then run a number of hypothetical scenarios, perhaps one in which the coronavirus pandemic subsides in three months and clothing goes back to regular demand levels. The model could, at the least, give the company an idea of the steps it would need to take if that were to happen.
This all requires a lot of computer power, along with human intuition, but businesses can, at least partially, predict the unpredictable and respond “hourly or literally daily to changes,” Waid said.
Static models that simply take in old data and use it to predict the future, however, aren’t much help. If you’re using one of these models now, Waid said, “you’re in trouble.”
Bill WaidGeneral manager of decision management, FICO
“If you’re predicting forward based on what happened three months ago, you’re in a different world,” he said. “This on-update is impossible unless you have machine learning models that can very quickly crunch those numbers.”
Beyond machine learning, companies are also turning to virtual assistants and chatbots to handle supply chain needs, said Ram Menon, CEO of virtual assistant vendor Avaamo.
“All supply chains are under severe stress,” he said. “Planes don’t fly; warehouses are unable to ship goods.”
Menon said he is seeing customers increasingly look to AI and automation to handle front-end issues, such as questions about order status or when invoices need to be paid.
Future of AI for supply chain
Now, relatively few companies other than giant manufacturers with the largest supply chains such as Coca-Cola and Proctor and Gamble are using agile machine learning and AI in the supply chain, Waid noted.
Those companies “are used to the fact where there is fluidity in their supply chain,” Waid said.
Yet, the coronavirus pandemic may force enterprises to use more AI in a variety of departments, including the supply chain.
The crisis showed how many organizations were unprepared for something a health crisis of this magnitude, Waid said.
“I do believe when we get to the other end of this, people are going to say, ‘I’m not going to get caught in that again,'” he said.
Ronen Lazar, CEO and co-founder of INTURN, a vendor of excess inventory management software, shared Waid’s sentiments.
Volatility in the supply chain, for many industries now, is outside of the normal, predictable levels, Lazar said. Companies are realizing now that they need more digital, adaptable technologies to help manage their supply chains, he said, adding that INTURN has been busier over the last few weeks than it had been in months.
“Sometimes innovation is driven by necessity. This is one of those situations,” Lazar said.
Avaamo, too, has been busier, Menon said. In addition to releasing a free virtual assistant that can answer dozens of questions about the coronavirus, Avaamo customers have asked the vendor to provide them with more virtual assistant technology during the pandemic.
Enterprise are plunging into AI and automation, turning to the technologies out of necessity during the pandemic and likely sticking with them afterwards, Menon said.
“Once companies cross the chasm, they never come back,” he continued. “This is automation’s big day.”