(Part 2)Warehouse Orchestration - PepsiCo’s innovative approach to transforming its warehouse operations

Warehouse Orchestration - PepsiCo’s innovative approach to transforming its warehouse operations (Part 2)

Keith Moore, CEO of AutoScheduler.AI, and Peter Hall, Senior Manager at PepsiCo, continue their discussion of Warehouse Orchestration in this second blog post. Don’t forget to read Part 1 first.

How does a warehouse orchestration project begin?

First, reformat your team to focus on specific areas of the operation:
  • Site design: How do we want the product to flow - from production through the warehouse to a trailer? What types of automation do we want to add?
  • What different technologies can we add?
  • WMS: How to make the WMS smarter and more efficient with how products are stored and moved?
  • People: What are their capabilities? How are we training the team on best practices in every aspect of warehousing? What key reports must we receive every day, every hour?
  • Intelligent warehouse: How can we bring in new technologies? What functions need to change? What technologies should we implement?
Define the data
  • What key data information is needed from the systems?
  • How can we get the data to a centralized location? What tools is the data going to use to feed into the centralized location?
  • How do we transform the data into a usable plan to orchestrate the work?
  • Start a pilot project
  • Create a pilot site and learn from it to understand what is missing, how it can be improved, and how to integrate the pilot into current systems.
  • How do we scale the pilot to full deployment? Deploy the warehouse orchestration system
  • Ensure teams act quickly to grow even faster in a way that can impact the network from a cost perspective.

As this project has evolved, what are you doing better than before, and what are you doing completely differently?

In our traditional warehouse, we would have a dispatch log or a load sheet that says these are all the loads for the day. We would go line by line and rely entirely on the warehouse's people to research the loads. Do I have the right product? Do I have the right people? Where is the product, what type of load is this? Does it require special loading conditions? Or can this go through automation? Does this need to be transferred to another warehouse area before it gets loaded?

Now, the orchestration platform just tells us what to do, such as starting this load at 5 p.m. and finishing it around 9 p.m., requiring X amount of effort to load it, and we're expected to load it at 100% or 98%, and then we can make faster and more informed decisions.

The orchestration platform uses AI and advanced mathematics to tell the team how to run the site effectively. It captures all the tribal knowledge and accounts for the constraints that exist with product flow. However, the company needs to undergo a change management routine to know how to best look at a screen, make a decision, save time, and run a more efficient site. Rolling out an AI solution in a network requires a lot of change management.