Most companies involved in industrial processes, or manufacturing have at least started their digital transformation, while other companies are well on their way to full digitalization. Where do you focus… where do you go once digital is operationalized?  Georgia-Pacific, believes that continuous transformation is an indispensable ongoing aspect of future success, and Cognitive Artificial Intelligence is turning out to play an important role.

We started the transformation toward Industry 4.0 about seven years ago.  It wasn’t something that we had 100 percent planned on, it was thrust in our general direction by necessity. Georgia-Pacific (GP) is one of the world’s leading makers of tissue, pulp, packaging, building products, and related chemicals. In 2015, the company’s innovation team was performing an in-house review of the operational model and noted that, at the time, a large portion of the workforce would be retiring around the year 2025. The issue wasn’t the sheer volume of potential retirees, instead, it was how to back-fill those vacated roles. We have a tremendous number of long-term loyal employees, and many of them have been with us for 25+ years. The question we focused on was, how do you continue to be a world leading manufacturer with people with little experience. 

"AI as a coworker will enable us to increase the speed at which we as humans conduct business"

One solution GP followed was to automate as many knowledge activities as possible. That enabled our workforce to monitor many of our new processes instead of performing the procedures manually as we had done in the past. It wasn’t changing what our teams did, but instead, it changed how they got it done. As a result, our whole operations team became more monitor-centric rather than operator-centric.

As the company progressed through the transformation process, we discovered that automation by itself wasn’t going to be the complete answer. The transformation of 150 sites isn’t the easiest task to handle. We learned that changing one process ultimately required us to change another and so on and so forth. We resigned ourselves to the understanding that transformation means transforming everything. Once the team had built this digital foundation and transformed much of the operational team to monitor-centric roles, we gained clarity to a plethora of new opportunities. These were opportunities to apply advanced technologies to fundamentally change the way the company addresses problems. We started thinking with a digital mindset.

COVID-19 and the associated challenges of 2020 affirmed that the company was on the right digital path. We had developed new digital capabilities, and in doing so discovered that our manufacturing facilities had not only maintained their normal production rates but in most cases set new production records even with quarantines and social distancing protocols. The pandemic, along with its unprecedented customer demand and logistical constraints, illuminated more areas of opportunity for GP. What we found was a supply chain system that was heavily dependent on manual intervention, redundancy, and inefficiencies. We plotted a course to change all of that, to solve these problems through technology, and to reduce the impedances that are felt by our customers. In essence, we set out to become the easiest company for our customers to do business with.

To achieve this goal, GP is turning to a new type of Artificial Intelligence, one that doesn’t fit into a typically narrow AI solution. This new AI can understand our business dimensions. It knows what a strategy is, what a capability is, what functions are and how these dimensions inter-operate. The resulting engine is basically an approach of concepts and methods on how to think about solving problems.

This type of cognizant agent can best be described as an AI Business Companion. It accumulates the knowledge of the business domain with the goal of identifying and eliminating any impedance from the perspective of the customer. We also learned that along with reducing the system’s inherent noise, which is the variability in human judgment within our business processes, that we could also capture the knowledge of our organization as the basis of effective decision-making while mitigating the loss of experience through retirements.

Imagine, if you had access to a system that contained the knowledge of all your employees, what would you ask it? That is the question the GP team asked themselves when they started identifying improvements to their supply chain. We found a lot of smart people who were weighed down by the transactional nature of our system.

One issue was big data. It’s not only the inconsistency in data quality but also the feasibility of fixing the condition of that data. Another issue was that traditional data science models wouldn’t function as intended in this case because we didn’t have all the data sets to train the model appropriately, and we didn’t find a good fit for traditional data science use cases. That’s when we began to look outside of traditional AI tools to cognitive AI and to an engine that can be taught through language and then apply simple logic to improve our performance in many areas of the enterprise.  

GP sees the future of AI forging through two pathways. The first is AI as a tool or a point solution.  AI as a tool has the potential to help solve a lot of problems, but it does so in a narrowly focused way: identify the concern, define the data you need, and then find an appropriate AI tool to help solve the problem. The second path is as a decision companion or digital coworker that understands your business, this type of tool helps you discover issues and then gives you recommendations on how to solve the problem utilizing data science methodology that is already available. 

Many meaningful steps for humanity have involved some form of companion relationship. When humans existed as hunters and gathers and first domesticated wild canines: In the beginning, humans relied on their intellect and own sight for tracking and acquiring food. Dogs on the other hand relied more heavily other senses to find food, their far superior sense of smell and hearing. Both the human and the dog were separately adequate, but the combination of the two enabled a much more capable solution. That’s also the effect of a decision companion. We can reduce the variability in our judgment because the digital coworker helps us validate the decisions we make. The speed of business has already exceeded the limits of human biology. AI as a coworker will enable us to increase the speed at which we as humans conduct business. Nothing else has appeared that can keep the noise within our human judgment in check while creating more opportunities for us to focus on our comparative advantages of ingenuity and decision-support.