There’s No Easy Button for AI and ML

It’s easy to get excited about using artificial intelligence (AI) and machine learning (ML) to transform supply chain and logistics operations. Take a bunch of data, run it through a model, and uncover the hidden answers that take your business to the next level.

The truth is, it’s a lot harder than that. There’s no Easy button: Push this, get that. Businesses must work on building their AI and ML from the ground up.

Companies willing to give AI and ML a chance often start with technological solutions versus focusing on the outcome, execution, and maintenance of the solution. There’s a far more effective way. Here’s how to avoid getting off on the wrong foot with AI and ML.


Start from Output, Not Input

The intuitive way to begin an AI/ML project is by looking at the data in front of you. Those models need a lot of data to produce insights or predictions, and many businesses get caught up in the details of what goes into the model.

Instead, focus on the output. What problems or business objectives are you trying to solve? Defining the question you’re trying to answer often indicates whether AI and ML is the right tool for the job. Some of the most common applications are: streamlining operational efficiency, optimizing procurement to achieve better margins, reducing WIP or inventory levels, increasing inventory turns, reducing lead time, identifying supply chain risks and disruptions, improving ETAs for more consistent on-time delivery, and creating better overall customer experiences.

Predictive and Prescriptive

Two of the strongest use cases for AI/ML are predictive and prescriptive analytics. A predictive example would be foreseeing a spike in trucking capacity heading to Florida during produce season because of a favorable weather forecast. With that in mind, companies could take advantage of cheap headhauls to move their goods into Florida at that same time.

A prescriptive case takes that even further. Using AI and ML models, businesses can employ a dynamic routing guide (an automatically running, rule-based method of interrogating sourcing options for a shipment). When signals from the market indicate spot rates have fallen below established contract rates, the routing guide short-circuits the traditional contract-based, waterfall approach. When the market corrects itself, the guide reverts back to its standard format.

The dynamic element of the routing guide takes the underlying automation and makes it responsive to unfolding conditions and adaptive in being able to learn over time.

Get the Right People

Most businesses don’t have the right people in place to adequately execute adding AI and ML to their processes. The key is to have a business translator on your team. This is an individual who can translate the business need to AI technology specifications and can help oversee the actual adoption and implementation of the technology while helping the organization own the business outcome.

A Virtuous Circle

With the right data and right people in place, businesses can create a virtuous circle. Information from your analytics tells your systems what to do, which identifies where AI and ML makes the most sense for your company.

But be patient when chasing the latest shiny AI and ML object. If you don’t have the right people and the right data to be able to execute your plans, then you have an expensive research project that sits on the shelf.

Leave a Reply

Your email address will not be published. Required fields are marked *