During the 2018 NAMA Show, the Coca-Cola Company hosted at least one session that caused quite a buzz. On Wednesday, March 21, was the session, Artificial Intelligence For Business Growth: How The Coca-Cola Company Is Leveraging A New Species Of Thinking. It began with a few words from Scott Corley, vice president, on-premise, of the Coca-Cola Company, welcoming everyone to the session and coming to learn about how artificial intelligence can benefit the industry. He then handed over the presentation to Jason Hosking, CEO of HIVERY, an artificial intelligence company.
Hosking told attendees that the artificial intelligent (AI) revolution is already upon us. He argued that AI allows computers to beat many experts, from chess to early diagnosis of breast cancer. AI is "able to perform tasks normally requiring human intelligence," he said.
In his mind, there is confusion because of movies, and the fact that AI is such a broad term. AI covers the calculator in your phone to self-driving cars.
Why AI is needed
Companies leverage only 12 percent of their captured data. It's impossible to do more with vast data sets without technology, Hosking continued. The answer is AI, and humans. He used the example of the error rate of diagnosing breast cancer. The error rate of pathologists was 3.5 percent, according to Hosking's chart. The AI model had a 2.9 percent error rate. However, using both the pathologist and AI model in combination produced a 0.5 percent error rate, significantly lower than either alone.
AI for the industry can: "Anticipate and meet customer need, increase revenue and decrease costs," he said.
Reza Kasravi, Head of IT Strategy, Architecture & Operations, Coca-Cola North America - The Coca-Cola Company, talked about core components of AI including robotics and sensors, natural language processing, computer vision, machine learning, neural networks and more. Much of this already exists from Amazon's Alexa technology to chatbots on websites.
Matt Robards, Cofounder, Data Scientist of HIVERY, followed Kasravi with an explanation of how humans make decisions. Many of the biases in our decision making is trying to make the option or decision simple, a tendency to stereotype or use hydrobolic discounting, which is favoring smaller, short term rewards over long term, larger rewards. HIVERY strives to move past that with data and predictive analysis. "The [HIVERY] cloud-based vending analytics can leverage existing data, can identify opportunities per vender basis and can predict what will happen if a vender made a change," explained Robards. In tests, the vending analytics showed a 4 percent increase in revenues while similar, traditionally maintained machines dropped 7 percent. It also improved the number of days the machine could go before needing service.
There is also a version for coolers that is called "Planogram Analytics," according to Robards. It can create a planogram for up to eight door cooler sets, share data and is a precursor to bringing the analytics in from vending.
Real-world test
Ed DeFraine, vice president food service and on-premise at Reyes Coca-Cola Bottling, rounded out the session by talking about his experience with the HIVERY analytics program. DeFraine described how his team wanted to get deeper into the specific patrons that were using their vending machines.
"We had a specific product set per type of location," he said. Operations, with the drivers input, did better than standard machines, but DeFraine knew it could be better. His team looked at the trial Coke was doing already with HIVERY in Las Vegas, NV area, and decided they wanted to try it in a pilot program in Sacramento, CA. There was a lot of discussion about how to measure the optimizing of the machines. How much could the revenue increase? What about the fill rates? How high was too high?
Following the trial, DeFraine admits to being happy with the success. "We were impressed," he said. "Analytics has become a way of life for us, and it's been readily adopted, unlike other technologies." It resulted in 15 percent fewer restocking trips and 6 percent in additional revenues.
Other benefits DeFraine mentioned was how it helped get clean data, and keep it clean. The optimization motivates everyone to maintain the data. It also changed the culture. Drivers will change the product set without question now because it's based on data and proved with revenues and trips. Plus, it's easy to use.
It came down to needing to improve the vending business, which led to AI, and DeFraine is happy it did.
During the 2018 NAMA Show, the Coca-Cola Company hosted at least one session that caused quite a buzz. On Wednesday, March 21, was the session, Artificial Intelligence For Business Growth: How The Coca-Cola Company Is Leveraging A New Species Of Thinking. It began with a few words from Scott Corley, vice president, on-premise, of the Coca-Cola Company, welcoming everyone to the session and coming to learn about how artificial intelligence can benefit the industry. He then handed over the presentation to Jason Hosking, CEO of HIVERY, an artificial intelligence company.
Hosking told attendees that the artificial intelligent (AI) revolution is already upon us. He argued that AI allows computers to beat many experts, from chess to early diagnosis of breast cancer. AI is "able to perform tasks normally requiring human intelligence," he said.
In his mind, there is confusion because of movies, and the fact that AI is such a broad term. AI covers the calculator in your phone to self-driving cars.
Why AI is needed
Companies leverage only 12 percent of their captured data. It's impossible to do more with vast data sets without technology, Hosking continued. The answer is AI, and humans. He used the example of the error rate of diagnosing breast cancer. The error rate of pathologists was 3.5 percent, according to Hosking's chart. The AI model had a 2.9 percent error rate. However, using both the pathologist and AI model in combination produced a 0.5 percent error rate, significantly lower than either alone.
AI for the industry can: "Anticipate and meet customer need, increase revenue and decrease costs," he said.
Reza Kasravi, Head of IT Strategy, Architecture & Operations, Coca-Cola North America - The Coca-Cola Company, talked about core components of AI including robotics and sensors, natural language processing, computer vision, machine learning, neural networks and more. Much of this already exists from Amazon's Alexa technology to chatbots on websites.
Matt Robards, Cofounder, Data Scientist of HIVERY, followed Kasravi with an explanation of how humans make decisions. Many of the biases in our decision making is trying to make the option or decision simple, a tendency to stereotype or use hydrobolic discounting, which is favoring smaller, short term rewards over long term, larger rewards. HIVERY strives to move past that with data and predictive analysis. "The [HIVERY] cloud-based vending analytics can leverage existing data, can identify opportunities per vender basis and can predict what will happen if a vender made a change," explained Robards. In tests, the vending analytics showed a 4 percent increase in revenues while similar, traditionally maintained machines dropped 7 percent. It also improved the number of days the machine could go before needing service.
There is also a version for coolers that is called "Planogram Analytics," according to Robards. It can create a planogram for up to eight door cooler sets, share data and is a precursor to bringing the analytics in from vending.
Real-world test
Ed DeFraine, vice president food service and on-premise at Reyes Coca-Cola Bottling, rounded out the session by talking about his experience with the HIVERY analytics program. DeFraine described how his team wanted to get deeper into the specific patrons that were using their vending machines.
"We had a specific product set per type of location," he said. Operations, with the drivers input, did better than standard machines, but DeFraine knew it could be better. His team looked at the trial Coke was doing already with HIVERY in Las Vegas, NV area, and decided they wanted to try it in a pilot program in Sacramento, CA. There was a lot of discussion about how to measure the optimizing of the machines. How much could the revenue increase? What about the fill rates? How high was too high?
Following the trial, DeFraine admits to being happy with the success. "We were impressed," he said. "Analytics has become a way of life for us, and it's been readily adopted, unlike other technologies." It resulted in 15 percent fewer restocking trips and 6 percent in additional revenues.
Other benefits DeFraine mentioned was how it helped get clean data, and keep it clean. The optimization motivates everyone to maintain the data. It also changed the culture. Drivers will change the product set without question now because it's based on data and proved with revenues and trips. Plus, it's easy to use.
It came down to needing to improve the vending business, which led to AI, and DeFraine is happy it did.
