Georgia Pacific Shows How AI and Generative AI Can Both Augment Operations and Accelerate Skills Adoption

Subscribe To Download This Insight

By Michael Larner | 3Q 2024 | IN-7443

Georgia Pacific demonstrates a method to avoid duplication of effort when it comes to solving issues around unplanned downtime, but also how generative Artificial Intelligence (AI) can be utilized to enable junior employees to acquire new skills and experience in a timely manner.

Registered users can unlock up to five pieces of premium content each month.

Log in or register to unlock this Insight.


Generative AI and AI Supporting Pulp and Paper Production


Like many other industrial and manufacturing firms, pulp and paper manufacturer Georgia Pacific is looking to utilize Artificial Intelligence (AI) and generative AI to improve its operational performance; in this case, by scaling Large Language Models (LLMs) and generative AI assistants across its 100 facilities to help employees identify and fix emerging issues.

The CSC Establishes and Scales Capabilities


One of the biggest challenges pulp and paper manufacturers face is the paper tearing or breaking, causing unplanned downtime (see ABI Research’s Digital Transformation in the Pulp and Paper Industry Update presentation PT-2747)) and manufacturers are looking to identify imminent issues from sensor data via data analytics and, increasingly, LLMs. In fact, Georgia Pacific aims to be able to predict a failure at 60 to 90 days’ notice. And similar to many other industrial and manufacturing firms, Georgia Pacific is conscious of losing expertise as members of their workforce head for retirement.

To prevent breakages, Georgia Pacific has partnered with KCF Technologies, Amazon Web Services (AWS), C3 AI, and SAS. First, the company has attached sensors from KCF Technologies to more than 25,000 pieces of equipment across its locations globally. Today, more than 60,000 sensors are constantly emitting data.

Georgia Pacific established a Collaboration and Support Center (CSC) and part of the organization’s remit is making sense of all the data. To achieve this, Georgia Pacific uses Amazon Kinesis (to send the data from the equipment in real time), Amazon Simple Storage Service (Amazon S3), which ingests and structures the data, and Amazon SageMaker to develop Machine Learning (ML) models. As a result, equipment operators and production engineers receive recommendations regarding optimizing the machine’s speed and suggested adjustments to avoid breaks and maintain quality levels.

Furthermore, specifically for avoiding unplanned downtime, Georgia Pacific uses C3 AI Reliability to analyze the data, seeking to look for evidence from, for example, vibration data for indications of looming part/equipment failure, as well as to generate work orders. Engineers can tweak the LLM models to monitor different parts and share them across the company; the establishment of the CSC aims to remove innovation silos.

Generative AI is used on the occasion that production engineers run into difficulties with the equipment or work processes. Georgia Pacific uses the SAS Viya solution, which provides generative AI assistants that can answer questions (based on the streams of data collected and processes in operation) regarding how to operate the equipment day to day and identify maintenance tasks and the part required to complete it. The benefits being improved include uptime, as well as upskilling junior engineers.

Putting in Solid Foundations for the Next Generation


For Georgia Pacific, “The goal of digital transformation is to make it so a three-year employee can operate as if they had about 20 years’ worth of experience.” - Roshan Shah, Vice President, CSC Operations, Georgia-Pacific.

Currently one of the main benefits of generative AI experienced by industrial and manufacturing firms has been the ability to retrieve accurate information from curated information, be that production data or equipment manuals (see ABI Research’s Generative AI Use Cases from Hannover Messe 2024 presentation PT-2911). For the person on the factory floor, context and accuracy are key. The solutions must provide information in a timely manner, but create rules to avoid being a distraction and be ignored.

To be successful, the AI and generative AI solutions need to be transparent and explainable to users both in terms of the outcomes and how the data and models have been created. This is critical when considering the consequences of decisions that are made such as scheduling fixes, procuring parts but not having excess inventory, and, in some cases, proactively halting production.

The CSC is an important step for Georgia Pacific to evaluate technologies, develop a technology stack, and be able to scale the innovations globally. But equally important is the company prioritizing the need to upskill junior employees quickly so they can step into the shoes of the veterans heading for retirement.


Companies Mentioned