Automation, Analytics, and AI in Supply Chain

Author: Sangeeta Gadepalli

“Siri, play the ‘Supply Chain Rebel’ podcast”

“Playing ‘Supply Chain Rebel’ podcast”

Life has become so much more simpler with IoT devices and personal assistants available at our beck and call. You can ask Siri for the weather or ask Alexa to order dinner. But that’s on a personal level. What about on the grand scale with large corporations? How are they using these incredible machines to get ahead of the competition? 

We are currently living through the fourth industrial revolution. What does that mean? According to the World Economic Forum, the fourth industrial revolution is a digital revolution transforming lives. While the previous three dramatically improved production capabilities, the fourth is transcending incremental development, blurring the boundaries between humans and machines. The intersection of human intelligence and machine capabilities has drastically altered how we work. Companies are increasingly harnessing the incredible potential of computer science, data modeling, and algorithms to improve processes, become efficient, and enhance customer experience. The three As of the digital era - automation, analytics, and AI - are dictating the future of industries and functions, including supply chain. Here are some of my favorite innovations holding the promise of a more efficient supply chain future: 

  1. Automate tedious tasks 

We can all agree that repetitive tasks are monotonous and tedious. According to an article on mhlnews.com, “Businesses estimate they spend on average per week around 55 hours doing manual, paper-based processes and checks; 39 hours chasing invoice exceptions, discrepancies and errors and 23 hours responding to supplier inquiries.(1)”

There are 3 main areas that could be automated here:

  1. Suppliers typically have questions during the procurement process. If a chatbot can support these inquiries, it’ll free up a significant portion of time for the procurement professionals to work on other important tasks. 
  2. Proofreading contracts is important but imagine reading 20+ pages of “legal” talk. Boring, right? Someone give me some coffee (yawn)! But if algorithms were trained to proofread to ensure all contract terms were written out explicitly, then that reduces man hours and human errors. Maybe you can enjoy that coffee while you watch your beautiful machine execute a brilliant algorithm!
  3. The supplier selection process is tricky. You need to factor in so many variables and selecting the right one out of a 100 potentials is a challenge. But that’s what machines are for. Sifting through enormous piles of empirical data like order fulfillment rate, raw material quality, accuracy of delivery, and lead time, algorithms can help companies select suppliers based solely on performance factors, excluding any pre-existing biases. But this one may be a thing of the future. What do you think? 
  1. Machine Learning Algorithms

Remember when you were 9 and your parents let you set up a lemonade stand. What did you need? Ingredients to make the lemonade right? But how much does a batch of lemonade serve and how many people were you expecting to serve? Maybe you didn’t realize it then, but you were a supply chain planner. You got raw materials, converted them to the final product, and delivered it to potential customers. But before all of this you had to plan for the right demand and execute it. Maybe you first planned for just your block. But once the word got out of how delicious your lemonade was, now the whole neighborhood wants it and pretty soon it might be the whole town! This is a similar scenario to how businesses plan to scale up their supply chains. 

The supply chain planning process is crucial to effectively deliver value to customers while keeping costs down. A primary component of this process is demand forecasting. Traditional forecasting models use time periods as short as 3 months but in the digital age even 3 month old data is too old. Using qualitative techniques like the Delphi(2) method and quantitative methods like time regression models(3), businesses can predict future demand with some level of accuracy. However, with advanced methods like predictive analytics, businesses can use real-time data and market trends to assess demand. 

One of my favorite supply chains (and store, hello European fashion!) is Zara. Instead of shipping once a month or an as-needed basis, Zara ships small batches in two-week increments. This allows stores to rotate styles throughout the year and creates a hypothetical “limited stock” environment, otherwise known as “better get it before it’s gone” syndrome. So even if a particular style doesn’t go as well as planned during the season, each store only has a small quantity of it. This allows Zara to accurately forecast demand, adjust stock according to the demand, and hedge any risks against a fashion faux pas(4). 

  1. Automating manufacturing process

Henry Ford introduced the assembly line in his automobile factory. He studied the different steps in assembling a car and worked with experts to automate the entire process. Instead of taking 12 hours to build a car, it took under 3 hours(5). This was done in the early 1900s. Now imagine what we can do in the 21st century. 

Robots have long surpassed the ordinary, providing value in every area from packaging to tooling, freeing up humans to do less repetitive and dangerous tasks. Robots have increased productivity and reduced errors, improving overall margins in the value chain(6). By automating production lines, the company and its staff can redirect their efforts to focus on quality, customer service, and most importantly - innovation & growth. 

  1. Warehouse automation

Costco may make warehouses exciting with an endless stockpile of everything you’ll ever need, but an actual warehouse is quite mundane. While the essential function of a warehouse is uncontested, the actual activities at a warehouse are quite predictable - receiving, storing, retrieving, trucking, and monitoring to name a few. Warehouses are an essential component of the supply chain, storing inventory before it gets shipped off to distribution centers and retailers or wholesalers. Now imagine introducing automation and robots here. Automated guided vehicles, or AGVs, can do the heavy-lifting by helping warehouse staff receive and store packages. Collaborative robots (cobots) or automated storage and retrieval systems (AS/RS) can help warehouse staff find and retrieve the right packages reducing time and walking in the warehouse (sorry fitbit. I can’t do my 10,000 steps anymore). Working alongside their human counterparts, machines can improve efficiency significantly at warehouses. 

  1. Internet of Things

Imagine ordering pizza after an extremely intense Spin class. You’ve blown past starving and are currently in the ravenous mode. Wouldn’t you love to be able to know exactly what the pizza place is doing? Thanks to the internet we can do that now. We can see when our order is received, when the order is being prepared, and when it is out for delivery. Now throw in your Siri and your Alexa that can track so much from your takeout order status to the weather outside. These AI powered personal assistants have transformed visibility bridging the gap between the merchant and a singular customer. But what happens when you scale these innovations up to the business world?  

In 2017, Michelin, the tire manufacturing company, launched its own IoT powered tracking system. The low-cost solution was a sensor attached by a magnet to a container that transmitted data about the container’s structural integrity over the 0G network. This solution was highly effective in not only improving visibility but also improving KPIs like shipping lead time (reduced from 40 to 35 days) and the company’s accuracy of the estimated time of arrival (improved by 40%) (9). 

Supply chain visibility is an enormous challenge costing tremendously. If suppliers, manufacturers, warehouse managers, retailers, and wholesalers had access to real-time data, it could improve visibility amongst all the key players. Internet of Things (IoT) sensors could be installed to shipments, thus transmitting real-time data throughout the value chain to provide visibility and reduce the bullwhip effect(8). IoT sensors can send data about shipments such as temperature, package integrity, etc. and this information, can be used by the manufacturer to respond quickly to any unfavorable conditions. 

  1. Autonomous vehicles and GPS tracking systems

Every supply chain professional is aware of the challenges of the last-mile. Along with visibility the last-mile is an optimization problem. Planning routes help ensure the deliveries are faster and reduce fuel costs, but these routes need to be adjusted based on traffic conditions and other factors. Autonomous vehicles and GPS tracking systems can help track the packages and provide optimal routes. 

UPS’ proprietary navigation system, ORION, provides their drivers with detailed delivery and pickup routes and instructions, thus improving their efficiency and customer service. Using real-time information like traffic conditions, pickup and delivery lists, ORION optimizes the routes to be more cost-effective and reduce miles traveled(10). 

If you’ve made it this far, you probably have a question or two floating inside your head. 

Question 1: How far away are we from fully automating these tasks? 

When considering automation, companies need to factor in technical feasibility, cost of implementation and deployment, and the ROI. In addition to these measurable variables, companies must also consider the implications of automation. 

Question 2: Are robots really going to replace human labor?  

Perhaps not! Investing in upskilling the workforce to operate machines, interpret data and analyze outputs is crucial to avoid the inherent displacement that unnerves the general population from embracing technology. 

Overall, technology has made serious strides in improving workplace repetitive tasks, synchronizing teams across the globe, and enabling the workforce to focus on innovation. We are at an exciting precipice of change and growth for the supply chain. If the current trajectory is any evidence of the future (forecast pun), then the global supply chain is in for a revolutionary awakening and I for one am excited for the ride! 

Although these are just a few ways in which automation is being currently used, there are several other applications. Let me know where you think automation could take the supply chain function. 

Sources:

  1. MHL News: https://www.mhlnews.com/global-supply-chain/article/22054569/supply-chain-losing-hours-money-to-poor-financial-systems
  2. Delphi Technique - https://www.investopedia.com/terms/d/delphi-method.asp
  3. HBR - How to choose the right forecasting technique https://hbr.org/1971/07/how-to-choose-the-right-forecasting-technique
  4. Zara’s Supply Chain Management - https://www.ivalua.com/blog/supply-chain-management-zara/ 
  5. Assembly Line - https://www.history.com/this-day-in-history/fords-assembly-line-starts-rolling
  6. Robots in Manufacturing - https://www.acieta.com/why-robotic-automation/robotics-manufacturing/
  7. What is warehouse automation? - https://6river.com/what-is-warehouse-automation/
  8. Bullwhip effect - http://www.cbpp.uaa.alaska.edu/afef/bullwhip_effect_in_supply_ch.htm
  9. Michelin IoT - https://www.dcvelocity.com/articles/30912-michelin-keeps-tire-shipments-rolling-with-help-of-iot-trackers
UPS ORION - https://www.ups.com/us/en/services/knowledge-center/article.page?kid=aa3710c2
>