How AI helps Coca-Cola boost supply chain procurement

According to Brett Fultz, director of global analytics, global procurement and supply chain at Coca-Cola, artificial intelligence (AI) and machine learning tools have become indispensable in order to boost procurement and sourcing efforts at the Atlanta-based global beverage leader. 

“Previously, we were the masters of Excel and email,” he told VentureBeat. “When I first started at Coca-Cola 10 years ago, we didn’t have any type of AI or advanced procurement or sourcing tools available.” Now, however, leveraging AI has provided great improvement not only for procurement across more than 50 categories, but also for supplier communication, governance and process, he said. 

For any company making or selling goods, procurement and sourcing are integral functions of supply chain management. Sourcing, an early stage of the procurement process, is about identifying and assessing potential suppliers of goods or services, negotiating terms and selecting vendors. Procurement goes further, however, and is about obtaining supplies and payment from suppliers that compete for business by submitting bids and negotiating contracts. 

But challenges abound in a supply chain landscape filled with disruptions and risks – from issues related to the COVID-19 pandemic and the war in Ukraine to climate change. For procurement leaders, that has become a huge hurdle: according to Deloitte’s Global 2021 Chief Procurement Officer Survey, the pandemic has intensified the challenge of detecting, measuring and managing risk. And the survey found that very few CPOs (18%) were formally tracking the risks that existed in their direct (tier 1) supplier base, and only 15% had visibility beyond that.

AI, however, can be a game changer: The Deloitte survey found that high-performing procurement organizations were 18 times more likely to have fully deployed AI/cognitive capabilities. 

AI can help boost decision-making amid procurement chaos

The current focus in procurement is one of “firefighting” – that is, daily efforts to stay on top of and ahead of the next seemingly unforecastable disruption, said Shanton Wilcox, U.S. manufacturing lead at PA Consulting. Today’s companies are faced with availability issues, wildly fluctuating commodity prices, transportation challenges and increased costs, he explained. 

“AI solutions help inform decisions in a seemingly chaotic business environment,” he said. Artificial intelligence can be in everything from the development of scenarios and modeling potential scenario impacts and decisions, to developing AI models that forecast prices and help balance the trade-off between current and future prices, he added.

Traditional models of procurement and sourcing systems that previously brought organizations success have become obsolete and strained, said Kushal Nahata, CEO and cofounder of FarEye, a supply chain management company that “leverages real-time AI insights and visibility” for companies including Domino’s and DHL.

AI has the capability to “unearth insights and patterns in the supply chain to find correlations that escape the purview of the human mind,” he explained. For example, determining the future availability of resources in an organization can calculate the approximate timeframe of deliveries. Historical data can be used to understand the probable demand in the future and accordingly make relevant decisions. 

However, AI is a fledgling area for many procurement departments, said Brian Prantil, senior vice president at Insight Sourcing Group. “The biggest area we have seen companies leverage AI is related to spend classification to deliver insights and identify opportunities,” he said. Other areas include using AI to monitor and identify potential risk areas across the supply chain, automate simple tasks such as checking status of purchase orders, and scan and interpret contract documents. 

Keelvar takes on supply chain procurement challenges

One company looking to tackle the procurement problem head-on is Cork, Ireland-based Keelvar, which recently raised $24 million to fund its “intelligent sourcing” platform. It has developed sourcing bots that allow customers to examine direct and indirect procurement bid information from suppliers and then analyze multiple awarding scenarios based on those criteria and other constraints. Using Keelvar’s platform, they can also launch and run new bidding events. Algorithms sort through data on supply chain disruptions and vendors, cleansing it and extracting information before offering recommendations.

As a Ph.D. student in an AI research laboratory, Keelvar founder and CEO Alan Holland worked part-time for his parents’ business, which sold chemicals to companies like Coca-Cola. 

“They were frustrated because procurement processes were difficult for them to navigate,” he said. “There was a tendency to simplify things by awarding to a smaller number of vendors because it was so hard to manage.” More sophisticated tools were required, he explained, to help procurement teams manage the inflow of data and capture much richer data from the marketplace.

Coca-Cola uses AI for supply chain visibility

Coca-Cola began using Keelvar two-and-a-half years ago specifically for logistics procurement, and now has incorporated it into its software for direct materials in packaging and ingredients. 

Fultz pointed out that his global procurement group controls over $25 billion in spend across over 50 categories, and certain categories have hundreds of suppliers bidding for Coca-Cola’s business. “With that comes all the pain points in the process of being able to standardize our data, standardize our process so that we’re able to move quickly for a sound sourcing recommendation,” he said. “What Keelvar has allowed us to do is help simplify that process.”

The most significant value, he added, has come in relation to matrix style pricing to parse out material costs – including raw materials, conversion and logistics costs – as well as automating the bid cleansing process. 

For example, “in a category like shrink and stretch film we would have well over 200 suppliers invited to an event and at the end of each round we would have all hands on deck to cleanse and validate these bids,” he said. “This tool has allowed us to streamline that process so that we’re essentially able to collect aggregated validated bids.” 

Other vendors offer solutions to procurement challenges

Many other, larger, AI-driven procurement solutions are in this space, such as Coupa and Jaggaer. Both of these vendors are leaders in the procure-to-pay space, which Gartner’s 2021 Procure-to-Pay Magic Quadrant, defines as “integrated solutions with automated workflows to request, procure, receive, and pay for goods and services across an enterprise.”  

Coupa emphasizes that it offers broader products and services than “niche, point solutions” such as Keelvar. “From sourcing to contract management to procure-to-pay, Coupa constantly measures supplier risk and performance,” said Dr. Madhav Durbha, VP of supply chain strategy at Coupa. “Through our supplier risk and performance management solutions, we can parse through a variety of internal and external data to risk score suppliers.” However, flagging a supplier as risky alone is not sufficient, he added. “Through the use of advanced optimization techniques on a digital twin model of the physical supply chain, we can run a variety of scenarios and compare the risk scores along with other KPIs.” 

He also pointed to Coupa’s Community.ai, which is anonymized, aggregated and analyzed data that combines AI and human connections. “Currently we have over $3.3 trillion of cumulative spend under management, which is a treasure trove of data that we leverage with our customers’ permission to abstract out insights that benefit the broader community through AI-powered algorithms,” he said. 

Stuttgart-based Makersite’s founder Neil D’Souza says complexities have increased drastically for procurement and sourcing since the beginning of 2020. While supply chain risk has always been a part of procurement and sourcing, he said, “Covid and geopolitical issues have made it harder and more important to master than ever, while visibility issues of …. suppliers pose large business risks,” he said. “Additionally, with rising regulations, and enterprises openly communicating NetZero goals, procurement is now in the driver’s seat to reduce emissions from the supply chain.” 

AI enables procurement and sourcing to tackle challenges at a massive scale, he said.  Makersite’s digital twin technology, he explained, delivers a “transparent and supplier-specific view of product supply chains and insights into supply chain deep tiers, associate risks, and emissions from purchased goods for manufacturing suppliers.”  

Going forward with AI for supply chain procurement

Investing in AI to boost procurement and sourcing will become a priority investment area in the near future, said Wilcox, due to complex challenges around business volatility and increased labor mobility. “Long gone are piece price comparisons – decisions now include aspects like flex capacity / responsiveness, initial sustainability interests, transportation options and resulting costs,” he said. 

Over the long term, procurement departments should consider investing in AI tools that will improve decision-making by leveraging internal and external data through anomaly detection, predictive analytics and advanced shoulder costing, said Prantil. That said, before investing, companies should evaluate their use cases and examine their data architecture to understand how they can leverage these tools. “AI can certainly automate cleansing and classification, but the data needs to be captured so defining what is and is not available is the first step,” he said. 

Data is absolutely the biggest challenge when it comes to implementing AI for strategic procurement, Prantil continued. “Data exists in purchase orders, accounts payable systems, contracts, specifications, drawings, vendor masters, spreadsheets, etc., and isn’t necessarily connected electronically, stored centrally or always very robust.” 

The future will bring an acceleration of AI and software systems in the procurement and sourcing space, predicted Durbha. They will, he said, “evolve into broader ecosystems and bring together data, algorithmic engines and capabilities to solve specific procurement and sourcing challenges.” 

Coca-Cola builds on procurement success using AI

Keelvar’s Holland said that in the future, the company will focus on integrations between multiple systems and best-of-breed vendors pulling in data from other sources. 

“There are great supply chain visibility tools so you can see where your containers and trucks are, and there are also providers offering sustainability data,” he said. “We feel it’s important that we start integrating more of that data in our systems – which is a trend that I think will gain momentum and will accelerate further in the next 2-3 years.” 

For Coca-Cola, there is still room to grow when it comes to improving supply chain procurement and sourcing, said Fultz. 

“In the logistics space we’re very mature, and now it’s really about bringing the rest of our categories up to that level,” he said. “Across the board, in our direct and commodity categories, we’ve had really great success. So it’s just really being able to build on that success and try new things, versus just the standard way of operating.”