Artificial Intelligence is Redefining Food & Beverage Manufacturing

Of all the disruptive digital technologies being deployed in food and beverage manufacturing, none is more controversial than Artificial Intelligence (AI). Robotics, Machine Learning, Predictive Analytics and AI are evolving from science fiction to reality, seemingly, overnight.

Of all the disruptive digital technologies being deployed in food and beverage manufacturing, none is more controversial than Artificial Intelligence (AI). Robotics, Machine Learning, Predictive Analytics and AI are evolving from science fiction to reality, seemingly, overnight. Companies are reevaluating their processes — and the workforce — and making decisions that will drastically impact the job market in food and beverage manufacturing.  

Today, AI is particularly valuable in automating and streamlining the complete manufacturing ecosystem, from speeding operations and automating routine decisions, to advising on executive-level decisions. It is also fueling a heated debate as to whether factories should ultimately rely on AI-driven machines and robotics, which could eliminate the majority of manufacturing jobs.  

Fear of the unknown will undoubtedly impede progress, which is why proactive manufacturers must understand the developments around AI and the modern application for the technology. While this is an emotionally charged topic, managers should delve into the facts and reach their own conclusions about the pros and cons of AI technologies and decide how AI — and other disruptive technologies — should be deployed in their enterprise. Although it can be intimidating to some, AI is an exciting technology that is here to stay. As it continues to rapidly evolve, companies must keep a pulse on the ways it will affect all aspects of the manufacturing industry.

The Current State of AI Technology

Most industry pundits seem to agree that AI is going to have a major impact on businesses. In fact, Gartner predicts, “By 2020, the number of users of modern business intelligence and analytics platforms that are differentiated by augmented data discovery capabilities will grow at twice the rate — and deliver twice the business value — of those that are not.”

Such optimistic thinking is helping to drive the large-scale adoption of AI technologies. AI is being embedded within many applications, from ERP solutions with built-in Business Intelligence capabilities, to Networked Supply Chains using Predictive Analytics. Consumers will start to experience the benefits of AI in many applications we have come to take for granted, such as e-commerce sites which recommend next purchases, or music streaming sites which suggest music for our playlists.

How AI is Being Applied in Food and Beverage Manufacturing

Facial Recognition. While many people associate this application with Facebook and its capability to “recognize” people in images and tag them, it also has some real potential value in manufacturing. AI applications for recognizing images can be used for security clearance to help companies better control access to their facilities. Additionally, the same type of technology that recognizes when an unidentified individual is trying to enter your building can also recognize if a potato chip does not have the perfect golden color.

Error Detection. Banks have been using AI for years to determine when “atypical” charges are being made against your account, triggering a fraud alert. The same type of monitoring for anomalies can be used by manufacturers to flag orders or purchases which are outside of normal patterns and indicate a possible error. It can also be used to monitor for compliance with regulations, workforce safety and the safe handling of food, including the cross contamination of typical allergens, like peanuts. The ability to “intelligently monitor” for potential liabilities means modern software can act as a watchdog and alert any abnormalities that we humans should investigate.

AI and IoT. AI and the Internet of Things (IoT) work together to interpret data received from sensors and can recognize when action is needed. Sensors generate such vast amounts of data that it would be useless without the ability to aggregate, sort and identify the data points which are significant. Often, the system is asked to spot anomalies such as early warning signs that an asset may be failing or may require maintenance. AI is also used to determine the seriousness of the flagged data points — such as if a technician should be dispatched or the line should be shut down due to imminent danger to workers.   

Personal Assistants for the Workplace. One of the most exciting applications of AI in manufacturing is the Personal Assistant, which uses Natural Language Processing to interact with the user, answer questions, perform functions and provide recommendations based on data science. This means a foreman, technician, R&D technologist or quality control inspector can verbally inquire about relevant details — from batch ingredients, to inventory levels — without needing to type on a keyboard or handheld device.       

Machine Learning. AI can streamline complex decision-trees, apply data-based science and proceed with defined actions based on parameters learned from user input. Laws of science, probability, and logic — specific to your industry — can be applied to predict outcomes, from buying trends, to pricing projections, to the availability of raw resources. Humans create the rules and define parameters, including the company’s risk tolerance. Machines can “learn” from experience what data points are considered acceptable and which are “out of bounds” and need action. More data helps to refine projected outcomes and improve the accuracy of forecasts.

Predictive Analytics. Using BI solutions with AI built-in, managers can see highly accurate forecasts of the future and evaluate clearly defined sets of recommendations to make well-informed decisions based on scientific reasoning. AI technology recognizes patterns, expands the knowledge base and discovers cause-and-effect relationships, using such insights to project the likely outcome or the next data point in the trend’s curve.

Automation. Numerous routine or administrative tasks in a manufacturing plant can be automated, thanks to AI. Workflows can be established, which allow data points to trigger reactions, signal notices, create reports, flag instances for review, place re-orders, dispatch crews, reserve parts and update batch schedules or recipe formulations based on seasonal availability or the quality of ingredients — for example, the sweetness of fresh berries may mean less sugar should be added.

Obstacles to Overcome

The term AI can make workers cringe as they envision a future in which machines run themselves and robots methodically execute the activities needed without a human in sight. It is true that in some industries the need for manual labor is reduced, but overall humans will still play an essential role in operations — and that will likely be the case for decades. In the current world of manufacturing, skilled workers are more important than ever.

False assumptions and fear of AI technology can get in the way of making improvements and hinder the move toward modern applications. For instance, workers may assume that implementing new technology translates to reductions in headcount. Technicians, engineers, logistics experts, and data scientists may — wrongfully — assume that a career in manufacturing is short-term and soon to be replaced by technology. Such fears could deter the next generation of skilled workers from even considering jobs in the manufacturing sector.

According to a survey by Constellation Research, 80 percent of organizations say they do not currently possess the human capital necessary to implement AI projects, while only 14 percent of executives say that their current staff is sufficient. Moreover, 40 percent of these executives reported that they will need to make significant talent acquisitions to support their AI projects.

What’s Next?

IT professionals who are skilled in AI are in high demand and difficult to recruit. Some manufacturers are training or re-skilling existing workers to foster in-house AI expertise, while others are finding that turning to third party resources and software providers is the most practical way to acquire expertise. When a manufacturer deploys technology in the cloud, the host’s experts become, in effect, the manufacturer’s experts, continually updating the technology and providing both back-up and security.

Partnering with a technology provider is also a sound method to ensure the most current and effective uses of AI technology are adopted. Leveraging modern technologies, like AI, is critical to remaining relevant and competitive in today’s landscape. Analyzing the past is no longer enough. Food and Beverage manufacturers must also envision the future with confidence, so they can make proactive and agile decisions that leverage AI and the most powerful technologies currently available.

*Gartner, Magic Quadrant for Analytics and Business Intelligence Platforms, 26 February 2018

Mike Edgett is Industry & Solution Strategy Director, Process Manufacturing, at Infor