Designing Super Supply Chain AI Solutions

super ai

Every automated solution is not an Artificial Intelligence (AI) solution.

There is still not clarity among many senior executives on what type of solution should be considered an AI solution. The primary driver behind this confusion is salespeople from hundreds of solution vendors who have started selling their automation solutions as AI.

AI Solution functional specs

This was the functionalities view. In this article, I will share a framework that evaluates a solution as AI vs non-AI based on the value delivered on certain aspects.

The SUPER Framework [1]

The SUPER framework provides five guiding principles that are necessary for successful AI implementation. They serve as the basis for an AI road map on which strategy can be built. The five aspects indicated in the illustration below need to be addressed by your AI strategy for your project to succeed.

Image credit: DC and Marvel

Understanding SUPER framework from a Supply Chain perspective

Now we will explore the five aspects of the SUPER framework in a Supply Chain context to get a functional understanding.

Proposed Solution

Let us assume that your organization is planning to embark on creating a smart, dynamic vehicle route optimization solution that also addresses backhauls. We will now use the SUPER framework to define functionalities that will ensure that the resulting solution is truly an AI solution.


Speed in this context is increasing the velocity of work and reasoning. Will the solution increase the speed of work or decision making process? An AI-enabled solution should do that by:

  • Automatically re-optimizing based on real-time adjustments to the schedule, traffic, disruptions, etc. Increases the speed of the decision making exponentially if the AI algorithms handle these changes.
  • By looking at multiple data sources (say supply and demand), the algorithm can make instant optimized decision on backhaul pickups and dropoffs, thereby eliminating back and forth between humans to arrange a backhaul.


We know that AI can access vast amounts of data very quickly, revealing and interpreting that information to provide insights. In this context, our Smart routing solution will perform the following (in addition to the ones listed above):

  • Based on weather data, interprets that storms may disrupt deliveries on a specific day so plans routes to prepone the deliveries and adjusts all the schedules accordingly.
  • Based on data generated by sensors on the truck, algorithm figures that it can accommodate a backhaul on the list and adjusts the schedule accordingly.


By performance, in this context, we mean efficiency, which measures the effectiveness of the process. We are concerned mainly with how well the job, service or product is or has been performing.

  • The tool should impact some key measures pertaining to route optimization like: Total miles, Total time, Driver utilization, deadhead, wait time, customer SLA compliance, fleet utilization, total outbound spend, etc. It should improve these measures much more significantly than a conventional route optimization tool.


Experimentation in the AI era means allowing for faster interactive processes, meaning supporting minimum viable product (MVP) feedback loops; creating, testing and optimizing. In the context of our route optimization tool:

  • Continuously evaluating how the algorithm interprets and uses all the pieces of data and information in its planning and decision-making processes and keep tweaking (training) the algorithm to make it better and better.


This one is a no brainer. The AI solution must produce results that support customers, businesses and industries. In this case, at a high level, the tool should be able to provide better customer service for customers, reduced operating expenses for the company and an innovative approach to route optimization for the Industry.


[1]: Based on the SUPER framework suggested by Chris Duffey