data cloud and the connection with the software

To Unlock AI for Supply Chain Management, Leverage the Data Cloud

Five Key Plays for Supply Chain Success With AI

Executive Summary

In this post we’ll share 5 key steps for successfully implementing machine learning (ML) and artificial intelligence (AI) within your supply chain.

We Heard You…

We received a lot of spirited feedback when I posted AI is no longer a hot topic in supply chain management. Let’s take a moment to clarify a few things:

  • Yes, AI still gets considerable air-time in strategy conversations, especially around innovation
  • Yes, AI has made some inroads, especially within warehouse and distribution centers
  • And, yes, AI will absolutely play a more prominent role in supply chain over time… 

But so far, the situation “on the ground” suggests AI has been overhyped when it comes to practical applications. And there’s more than one culprit: from culture to process to execution, supply chain just hasn’t proven to be a very good seeding ground for AI.

However, things are changing — perhaps faster than some of us expected. The catalyst has been data cloud technology, which has quickly made strategic applications of AI in supply chain more viable. As we scan the environment for early movers in this space, we’ve extracted five keys that comprise a playbook for near-term success:

  1. A Good Problem: that is, worth solving, and solvable. Not a technology looking for a solution.
  2. Expertise: Real data competence (or at least knowing how to ask for help) is part of the success formula.
  3. Good data: Clean, real-time, reliable data is essential.
  4. Scope: Real world problems are discrete — scope expansion has to be controlled.
  5. Execution: Intelligence initiates, but solid execution and measures are essential to grow impact.

Let’s take a closer look at each of these keys:

A Problem Worth Solving

At the outset of any proposed AI project, ask yourself a simple question, “have we tried solving this before?” If the answer is “no,” then…how painful, really, is the problem? On the other hand, if the answer is “yes, multiple times,” then there’s a good chance you and your colleagues will agree it’s a problem worth getting after, and eventually getting right.

Expertise

No doubt, your IT team has a wide range of expertise. But if you’re like most firms, AI and Big Data aren’t necessarily part of their core skills. If you have a dedicated Data Team or Data Science team, you may be one of the lucky ones — but make sure they have capacity. These days, most data teams are swamped with IT-authored projects. 

In many cases, the right solution provider with the right implementation team can get you 90% of the way there, but make sure they’re crystal-clear about what they do, and what they expect you to do.

Good Data

It’s helpful to bear in mind that AI-based solutions get “smart” by recognizing relationships in the data. The more data, the more obvious the relationships, and the “smarter” the technology becomes. But if you ever wonder “why isn’t AI used more in supply chain,” the most likely culprit is…messy data. However, this is changing, fast. With the rise of the data clouds, we finally have means to extract the right supply chain data and make it useful. We’ve talked about the stampede to the data clouds before. Enterprise data is being moved to the cloud fast, and supply chain is next. We believe this macro-trend will pave the way to more frequent AI successes.

Scope Control

The old advice is good advice. Start small, test, learn, and grow. But you’ll want to stay connected to the idea of fixing a problem worth solving, and adopt a prove-as-you-go model. You don’t have to solve the whole problem at once — but you will have to show real progress, and results, to fund your next phases.

Execution

Continuous optimization is required for automated learning to “do its thing.”  This means making sure that you’re ready on the ground to execute on the insights you derive. Scope the data, decisions, and actions you’re looking to drive ahead of implementation. It’s worth noting how this idea refracts back-through our previous keys: particularly Problem Selection, Scope, and Data.

Yes, AI is a Change Management Challenge

One thing you’ve probably noticed as we’ve briefly run through this list — for all the bold new potential for AI and Machine Learning, implementation of these new technologies draws from well-established change management practices. 

For other recommendations around change management specific to our space, check out these best practices for change management in supply chain.  

We’re finishing up a fuller treatment of these five keys AI For Supply Chain, so watch this space for more details in the near future. We’ll dig in a little deeper for each key, and bring some real-world examples into the conversation.  See you then — we’re looking forward to it! 

David Blonski

David Blonski

Related Posts

person choosing path

Responding To Uncertainty

Four Critical Steps That Will Separate Supply Chain Leaders as Recession Looms In Part One of this blog post, we discussed in some detail how

Read More >

See how Enterprise Service Management can help your business