Machine Learning

Machine Learning: AB InBev Case Study

AB InBev is reducing manufacturing costs and improving the quality of its beer by using Google Cloud including machine learning.

Google Cloud Results

  • Increases the barrelage per run by 60% through longer filter runs.
  • Reduces costs of filtration.
  • Delivers the best possible beer taste through machine learning.
  • Stays competitive by adopting the latest tech for manufacturing.

What goes into refining the taste of beer? Malt, hops, rice, yeast, water — and ML.

Anheuser-Busch InBev (AB InBev) makes some of the world’s oldest and most popular beer brands — including Budweiser, Corona, and Stella Artois. But the global corporation is all about embracing the future and innovating to stay competitive. That’s how AB InBev came to partner with Pluto7, a technology solutions provider that uses Google Cloud services to improve operations at manufacturers and other companies.

Revolutionizing Beer Filtration with Machine Learning

Pluto7 developed a prototype solution that is enabling AB InBev to optimize the beer filtration process with much greater accuracy — reducing costs, increasing efficiencies, and perhaps most importantly for beer aficionados, ensuring taste. The Pluto7 solution combines TensorFlowCloud Machine Learning EngineCloud SQL, and BigQuery.

  • Working with Pluto7, the AB InBev team evaluated six months of manufacturing data from its Newark, New Jersey brewery.
  • The data was fed into a TensorFlow ML engine run through Google Cloud.

“We studied the potential impact variables and matched them relentlessly against previous data from the brewery’s filtration process, taking advantage of the supersonic speed of Cloud Machine Learning Engine and its capability to analyze at a far swifter and comprehensive rate,” says Adam Spunberg, Global Director, Tech Explore, the AB InBev in-house technology group and part of the company’s Tech Supply Program.

“Through that process, we honed in on more than 50 specific parameters that displayed potential predictive ability, and we leveraged the highest coefficients among those to revolutionize our system.”

“This project never could have been so successful without the extraordinary cooperation and relationship with Pluto7 and Google,” Adam says. “From day one, it felt like we were one team, and that camaraderie grew, from the initial model in the Makeathon, to the 60 percent improvements we saw per run in barrelage.”

  • The success that AB InBev has achieved with Pluto7’s help and Google technology is also getting other people’s attention.
  • AB InBev was recently showcased at the 2018 Google Cloud Next conference in San Francisco and selected as a Finalist for the Supply Chain Breakthrough of the Year awards at Gartner’s 2019 SCM World conference.

Machine Learning: Optimizing Beer Filtration

  • A top priority at AB InBev is optimizing the K Filter that kicks in at the end of beer brewing, right before packaging.
  • The filtration process removes any remaining yeast, proteins, or other elements to achieve the best, crispest beer taste and meet brand-required turbidity levels.
  • Beer filtration is a complicated process with many unpredictable variables, such as the fluid’s turbidity coming into and going out of the filter as well as regulating pressure in and across the filter.
  • The technology long in place to manipulate the variables is capable of only basic logic, using meters to monitor and react to adverse conditions, such as a change in pressure.
  • Adding ML and artificial intelligence (AI), the company can leverage a much larger dataset to better predict and prevent potential issues during filtration.
  • Consequently, an ML-enabled filtration process could deliver great-tasting beer with significant cost efficiencies. Recognizing this potential set AB InBev on a search for the best partnership and provider that would enable the company to deploy ML quickly and effectively.

A Month-Long Makeathon

The AB InBev Tech Explore team had been actively exploring how AI and ML could solve some of the company’s top, ongoing business challenges. In late 2017, the incubator held a Makeathon, inviting providers of ML and AI solutions to participate. During the month-long bake-off, 14 solutions providers developed Proofs of Concept (POCs) that addressed the challenges that AB InBev needed to solve. Each technology provider was encouraged to develop a solution to 1 of the 12 use cases that would best showcase their capabilities.

Tech Explore collaborated with Google for the event. During the Makeathon, multiple Google Cloud partners developed solutions for the K Filter challenge, knowing it was among three priorities identified at AB InBev. Developing chatbots and improving video analytics were the other two.

Currently, AB InBev is working with Pluto7 to determine the best way to scale the solution to multiple brewery locations. Over time, the plan is to deploy the solution at 12 AB InBev breweries in the United States and at all K Filter and similar filtration systems globally.

“The value of working and the partnership with Pluto7 is that they’re scalable with the Google Cloud — so the learnings we get from Newark, we can scale across breweries in 126 countries,” says Benjamin Lavoie, Global Director of Technology at AB InBev. “The artificial intelligence of Pluto7 and Google Cloud is a part of this broader vision to transform how we make beer, and deliver it faster at more consistently higher quality to our valued customer.”

Endless Possibilities

In many ways, beer-making relies on traditional manufacturing processes. But AB InBev is open to shaking things up.

  • “AB InBev is thinking a year or two ahead of its industry,” says Manjunath Devadas, Founder & CEO, Pluto7. “They’re serious about exploring how technology can improve the manufacturing process in ways never before possible.”
  • “Based on the success in the K Filter project, we have already identified several more AI ML opportunities that are perfect fits for this partnership,” adds Adam.

“The K Filter optimization project was so critical and valuable because it showed the business how transformative AI can be and won us support and buy-in going forward.”

  • Adds Pat Fagan, Senior Quality Manager for AB InBev: “Knowing that whenever someone opens up a Budweiser, machine learning played a role in how it tastes? That’s cool.”

Request for a Call

Collaborate with the best in the industry. Let’s talk and get your project moving.


Contact Us For Questions

Thank You for Subscribing

Chatbot Dialogflow CX Instructions