How Some Food Companies Lean on Artificial Intelligence to Address Supply Chain Bottlenecks

How can technology that leverages artificial intelligence help companies reduce operational wastes? Software developer ThroughPut’s COO Seth Page answers this question, looking specifically at enterprises in the food space –companies and their upstream suppliers, their downstream shippers, and retailers. He discusses the lack of visibility into the scope of a massive problem: food loss, and he talks of its consequences. Page  gives his take on how companies can make more systematic changes to use their resources more efficiently.

Waste360: Please describe Throughput’s platform and how it supports whole supply chains in reducing their waste. And explain the difference between responding to “actual demand” vs estimated demand.

Page: ThroughPut is a supply chain artificial intelligence (AI) software platform that empowers companies to leverage their existing data to reduce their operational wastes and achieve better business, financial, operational, and sustainability outcomes. 

The software is hosted on a secure cloud, connecting disparate data systems to analyze, visualize, and provide actionable recommendations for better supply chain results. This is done within clients’ budget parameters and without expensive consultants, digital strategists, and multi-year “transformations.

ThroughPut leverages clients’ existing data systems with its supply chain AI to reorient material replenishment and deliveries based on dynamic lead-time changes, real operational consumption, and actual demand. In contrast, current planning systems typically assume fixed lead times and operate on estimated forecasts, which are inaccurate due to dated assumptions, supply chain disruptions, and bottleneck limitations. 

ThroughPut leverages any data sets taken from procurement, ordering, and planning systems. The software reprioritizes what, when, where, and how much to ship, as well as the best transportation mediums, optimal routes, and most profitable configurations.

Waste360: What are some of the biggest bottlenecks in the food supply chain?

Page: All fragments of the food supply chain face production bottlenecks, but imported perishable goods especially contend with unique challenges. Food products such as exotic produce have been hit hard due to labor shortages in developing countries that supply the food. 
One main contributing factor to food waste issues is a lack of visibility into the scope of food loss in all stages of the value chain. This lack of awareness is widespread among manufacturers, consumers, food service providers, restaurants, and others. Food waste it on the rise globally and increasing every day, so the food industry must look for innovative, digital-forward ways to combat this disastrous issue.

Waste360: Where are you seeing the most problems around food waste?

Page: Between 33%-50% of food produced globally is wasted. On top of that, 135 million people suffer from acute hunger, and the pandemic put an additional 130 million people at risk. The waste is happening along the whole value chain, but it’s most visible at production and consumption levels. The problem is especially noticeable at the production level in developing countries and at the consumption level in developed countries. 

Waste360: What are challenges with trying to address those problems? And how do you see technology helping?

Page: Companies in the food industry have historically tried to use big data solutions to solve supply chain issues. However, these solutions aren’t intelligent enough to overcome food waste issues, which is why artificial intelligence, machine learning, and IoT are so important. These tools help find trends and correlations across a huge number of variables and data from internal and external sources, offering an all-encompassing view into the supply chain bottlenecks that cause food waste and other issues. These technologies are a way to get out of the vicious circle of production and wastage, as they allow food companies to make more systematic changes in their supply chain to use their resources more efficiently.

Waste360: What experience does your team bring to the table?

Page: ThroughPut's team boasts many years of experience, including former field operators who managed planning and execution for Fortune 500 complex supply chains during economic downturns in warzone environments to maximize on-time-in-full, while keeping personnel safe and working alongside military and intelligence agencies.

Waste360: Can you speak to bottlenecks driven by COVID? How were/are you helping your clients deal with COVID-related supply chain issues?

Page: COVID-19 caused many supply chain bottlenecks, perhaps most noticeably in toilet paper. We saw empty shelves in almost every grocery store due to panic buying. However, we also saw empty shelves of meat, milk, flour, and other staples. Even when consumers didn’t specifically need these products, they panicked and scooped up as much as they could. This caused even more issues in an already-precarious food supply chain. It also showed just how easy it is for supply chains to break down.

From the first outbreaks of COVID-19, ThroughPut worked with clients in the food industry to immediately analyze their rapidly shifting true demand on a SKU, store, and routing level; match it to their actual capacity further upstream; identify bottlenecks in between; and optimize their supply chain network. 

Waste360: Where is most of the food waste occurring, and how well is existing technology doing to help deal with it?

Page: Humans have a waste problem. Of the $94 trillion global GDP, an estimated $25 trillion is operational waste endemic to supply chains. Much of this waste occurs before an end customer has a chance to see or purchase it. It’s avoidable waste leading to unsustainable business practices.

Contemporary technology solutions, however, still deal with static models, outdated technologies, poor assumptions, and simple analysis and visualizations. They don’t solve the big waste problems. Rather, they focus on point solutions in a near vacuum, versus system solutions that solve holistically across the entire end–to-end supply chain. 

Waste360: What organizations and industries use your technology?

Page: ThroughPut works with many businesses and organizations across numerous industries. Most pronounced during COVID-19 has been food, agriculture, beverages, and their containers; building materials and cement; finished goods; semiconductors and electronics; distribution and transportation; aerospace, defense, and automotive. 

Waste360: How is ThroughPut able to work for any industry?

Page: ThroughPut’s supply chain AI software suite optimizes across any organization that “makes” or “moves” things. The AI continually learns and applies those lessons and optimizations to new industries as it is exposed to them, thus freeing up domain experts to focus on what they do best, and not be human calculators.  

Other team leaders worked for Tesla, leading the back-end systems to launch all four electric vehicles, as well as tying Walmart’s backend systems to the web to create Walmart Online. 

Waste360: Why is AI technology so important in addressing supply chain issues? And what do you see as the smartest way to use it?

Page: AI is simply another advanced tool to take over the grunt-work of massive number-crunching and data analysis that is normally handled much slower by humans. By offloading the tedious tasks to AI, you free-up your in-house domain experts for more productive work, like making and implementing value-based decisions based on the outputs of AI-powered software. 

AI isn’t magic and if implemented properly, it should never be at the forefront of software. Rather, it should quietly analyze massive amounts of data, using tuned algorithms, to provide recommendations to humans. Humans can then interpret those recommendations and make decisions on what and how to implement them.

ThroughPut software handles hundreds of thousands of SKUs, coming from hundreds of upstream processors and sources, through dozens of distribution centers, hundreds of warehouses, thousands of stores, and tens of thousands of regular truck movements, representing billions of rows of data per client project. 

No human could, nor should, manipulate all that data to try and make sense of it. Let AI do what it’s good at — the data crunching and situational analysis. And let humans interpret the outcomes and put the best practices in place according to their own judgment.