3 Ways Our Partnership With Scale AI and Amii Are Benefiting Businesses
“We are we're really in service of one thing, which is decision making and decision support autonomously at scale at a distance.”
- Routeique CEO and Co-Founder Mike Allan
Key Takeaways:
- We’re partnering with AMII and Scale AI to create AI-powered supply chain solutions that enable faster, more accurate decision-making.
- These solutions will enhance demand forecasting, sequencing, and warehouse layouts.
- We’re currently partway through the project, and already seeing great results.
Project Background: The Benefits of AI-Driven Supply Chain Software
Our solutions can benefit our supply chain customers by allowing them to rapidly make decisions based on large volumes of information.
AI-driven supply chain software can take millions of data points from across a vast network in real-time, and analyze them–this is a task that is virtually impossible for a human to do.
Then, they can provide the relevant analysis or options to the human user to make the final decision. In other cases, the solution itself can make the decision, freeing up workers to allocate their time to other tasks that better build the business.
Our Process: Working With Leaders In The Machine Learning Space
Our team has been working in the AI and ML space for years, but our partnership with Scale AI and Amii is a chance to greatly expand this.
“Our route optimization system was built by our team and we have some pretty serious chops in that area,” says our CEO and Co-Founder Mike Allan, “But with Scale AI, were able to work with some of the rock stars of the AI machine learning industry, Dr. Matt Taylor [of AMII] and a number of other people like that that really bring an air of of the best of the best, best of breed technologies and and using the latest most modern tools like reinforcement learning and things like that.”
In addition to working with leading supply chain partners, we have been building out our internal team and upskilling existing team members in this area.
With our team currently currently progressing through this innovative project, read through our blog to learn three of the key ways it will benefit users.
Our Process: Working With Leaders In The Machine Learning Space
Demand Forecasting
The first use case for this project is demand forecasting. This enables businesses to carry the right amount of inventory. When you are carrying the right amount of inventory, you avoid problems associated with too little inventory, like shorted customer orders, as well as waste related to carrying too much, like expired products or needlessly taking up valuable warehouse space.
“I think what's what's really challenging and what we're working through right now is, you know, how do we not only provide you with a forecast, but contextualize that forecast for you in a way that you understand what it means, what it includes, our projected accuracy of of that demand forecasting,” says our VP Product, Bent Bawel, “...we're working with our partner Birkby Foods right now where we're looking at one of the particular streams of their buy-sell business to evaluate, you know, how can our models perform over a series of different products and SKUs and how does that compare to what they're doing right now in terms of their own expert who's kind of in there working on it? And we're trying to take some of those learnings from the expert and build them, building those ideas into the model. Again, we want the end result of this to be that it has the ability to work for a lot of different clients, and a lot of different product types. And so, you know, we want to take some of those things and bring them in.”
Our team has looked at a variety of different models, such as mathematical models, as well as reinforcement learning models, to see which performs best across different product lines. “Now we're in that phase where we're really just trying to tweak and glean a little bit more accuracy out of the models with each step,” adds Brent.
Sequencing
As mentioned by Mike, we have been working on sequencing in the form of Route Optimization for many years. Bringing this solution to warehousing yields many benefits, such as reducing travel time within warehouses. This can result in both more efficient picking, as well as reducing worker fatigue.
Within a warehouse, sequencing looks like a series of orders, provide it to a series of team members and have them most efficiently pick and pack the products.
However, our team is working on optimizing for factors other than just distance. “[The sequencing] part is working really well now,” says Brent. “The kind of interesting pieces, bringing in some of those other factors that are related to the picking and packing process. So that might be looking at some of the lot expiry-type things. It might be looking at weights or how they need to pack a pallet…”
Additional elements the team is looking at in terms of optimization relate to optimizing sequencing and storage not just within a single warehouse, but across multiple warehouses in a network.
Warehouse Layouts
With warehouse space at a premium in many markets, making effective use of storage space is absolutely essential. A well-optimized warehouse will deal with less congestion and will be able to store more of the right kind of goods where needed.
Our foray into space use optimization began when we started working with the Big Rock Brewery, team, who were seeking to optimize space in their warehouse. (Read the full case study here).
The initial use of this solution “...we've got the model working really well. And when I say really well, what's important for us is that “It's showing a diversity of results to the end user to be able to [...] see how different types of operations could look in a particular space.”
Spaces can be optimized for a number of things, such as unique pick faces, storage space, accessibility, and more.
The ideal layout can also be affected by factors such as:
- is the business going to block-store products?
- are the products in question ese fast-moving items or these slow-moving items?
- how many team members will be picking orders?
Brent adds, “And, you know, the model could produce anywhere from a couple of hundred to a couple thousand different layouts for them to choose from. And of course, we're not going to show them all thousands of these, but we're going to show them a filtered view of them, which allows them to see the diversity and make choices around how they actually operate their business to get to that perfect result for them.”
Find Out More About How Our Supply Chain Software Serves You
Want to learn more about our other innovations as our team continues to create new and exciting AI-driven supply chain software?
Head over to LinkedIn to follow us, or sign up for our newsletter to get the latest info right in your inbox.
Want to learn more on the topic of the three use cases, as well as answers to questions from business owners on topics like choosing the right AI solution for your needs, and more. Check out this full-length discussion panel moderated by AMII’s Adam Danyleyko, Product Owner, of Startup. Plus, stay tuned for our upcoming video series on this event!
