Getting Started with AI in Supply Chain Management (Lessons from the Past)

As I highlighted last week in “Negotiating With A Chatbot: A Walmart Procurement Case Study,” despite all the hype surrounding Artificial Intelligence (AI), relatively few companies are using AI in their supply chain or logistics operations today.

Simply put, we’re still in the very early adopter stage.

When it comes to adopting AI in supply chain management, are there lessons learned from the past to help guide companies today?

With this question in mind, I searched through our Talking Logistics archives and came across a post from November 2012 that offers some applicable advice.

At the time, most companies were still in the “early observer” stage when it came to using social networking tools for business processes other than marketing and customer service. GE, however, was an early adopter. In November 2012, MIT Sloan Management Review published an informative interview with Ron Utterbeck, who at the time was the CIO for GE Corporate and the Advanced Manufacturing Software Technology Center in Michigan (Ron is now the CIO at Eversource Energy). The article was about GE’s implementation of GE Colab, a social networking platform the company had launched in January 2012 and was being used by 115,000 employees worldwide at the time.

I encourage you to read the article for all the details, but below are some of my key takeaways that I believe also apply to getting started with AI in supply chain management.

(Note: I shared these same takeaways in my November 2012 post, but I edited some of the text to reference AI instead of social networking). 

Focus on power users, not business functions.

Investments in technology are often driven by the needs of a specific business function. A big obstacle many companies face with AI, however, is that they don’t know where to begin, where within the company to conduct the first pilot test. GE took a different perspective with social networking. Instead of looking for WHERE to begin, the company looked for WHO to begin with. “We rolled it out to our power users,” Utterbeck explained. “We didn’t focus on a function – the functionality [in the platform] is needed by every function. We really sought out the most experimental people in the different functions, and seeded [the network] with them and then got their feedback.”

Don’t wait for the “perfect” solution to get started.

First, the perfect solution doesn’t exist, and won’t ever exist, so stop waiting for it. And second, you don’t know what you don’t know — in other words, since AI in supply chain management is a new frontier, the only way to truly know what you need and want is to start the journey and see where it takes you. In GE’s case with social networking, the company started with a base product and “extended the heck out of it” based on feedback from power users. According to Utterbeck, “We launched [the platform] knowing that it was good enough to get people to start moving on it, and then we started to get feedback…and we incorporated that feedback into quick releases.”

In short, implementing an AI solution is not enough; you also need to have an easy process for users to provide feedback and ideas for new capabilities, and a process and resources to turn that feedback into new functionality quickly and frequently.

Don’t waste time coming up with an ROI.

One of the main reasons why companies aren’t using AI in supply chain management is the difficulty in quantifying its business value. The same was true a decade ago with social networking. However, that didn’t stop GE from getting started. “We haven’t tried to come up with an ROI,” said Utterbeck. “Haven’t wasted a moment’s notice even thinking about it.” The company tracked usage, adoption, how people were using the system, and what their connections were. Utterbeck goes on to say:

“The biggest thing about usage is that no one in this corporation has to use this platform to get their job done. It’s not a system that people have to go to, but people still come back every single day. They come back because it makes their job easier, because they’re getting value out of it. Going and spending money on ROI would be, honestly, in my opinion, just a waste of money because the true value of this is that people are coming back.”

Do you agree with these recommendations for getting started with Artificial Intelligence in supply chain management? What other lessons learned from the past would you include? Post a comment and share your perspective!