The North American paper industry is navigating a complex landscape. Rising raw material costs, increasingly stringent environmental regulations, and a widening skills gap are creating significant operational and financial pressures. Finding and retaining skilled workers in critical areas like production, maintenance, and logistics is becoming progressively challenging, hindering efficiency and limiting growth. To thrive in this environment, paper manufacturers across Canada, the United States, and Mexico must embrace innovative solutions, and Artificial Intelligence (AI) offers a powerful and increasingly accessible path forward.

The Rise of AI: A Transformative Force in Paper Manufacturing

Artificial intelligence is no longer a futuristic concept; it's a rapidly maturing technology that's transforming industries worldwide, and the paper sector is ripe for disruption. AI offers the potential to automate tasks, optimize intricate processes, enhance decision-making capabilities, and, crucially, address the growing skills gap that threatens the industry's long-term competitiveness. To understand the full scope of this potential, it's useful to consider different levels of AI implementation: general AI concepts, the targeted capabilities of enterprise-grade AI solutions like SAP Business AI, and the cutting-edge advancements of agent-based AI like SAP Joule.

Enterprise-Grade AI: Beyond Simple Automation, Delivering Real-World Impact

While the term "generative AI" might initially bring to mind tools that create text or images, enterprise-grade AI solutions, such as those offered within the SAP Business AI portfolio, go far beyond basic content generation. These systems are designed to integrate seamlessly with the core operations of complex businesses, like paper manufacturing. They function as highly specialized, industry-savvy "digital experts," understanding the unique challenges and opportunities within the paper industry.

These AI solutions are often embedded directly within existing enterprise resource planning (ERP) systems like SAP S/4HANA, providing contextual insights and recommendations within the familiar workflows that employees already use. This embedded intelligence is a key differentiator. Unlike generic AI tools, enterprise-grade AI is trained on vast amounts of industry-specific data, allowing it to understand the intricacies of invoice processing, supply chain dynamics, customer relationship management, and other critical processes unique to the paper industry. Furthermore, leading providers prioritize data security and regulatory compliance, ensuring that sensitive business information is protected – a paramount concern for any modern enterprise.

Combating the Skills Gap with Intelligent Automation and Support

One of the most significant benefits of enterprise-grade AI is its ability to mitigate the skills gap. These systems can translate complex procedures into easy-to-follow instructions, empowering less experienced employees to perform tasks that previously required specialized expertise. Imagine a new maintenance technician being guided through a complex repair with step-by-step instructions and visual aids generated by AI. This not only improves efficiency but also reduces the risk of errors and enhances on-the-job learning.

Beyond real-time support, AI-powered platforms can create personalized training modules, accelerating the upskilling process for existing workers. These tailored learning experiences can focus on specific skill gaps identified within a company, ensuring that training investments are targeted and yield maximum impact. Moreover, by automating routine, repetitive tasks like data entry and report generation, AI frees up skilled employees to focus on higher-value activities that require critical thinking, problem-solving, and strategic planning – precisely the areas where human expertise is most valuable.

AI also plays a crucial role in enhancing overall productivity and optimizing resource utilization. These systems can analyze vast amounts of historical and real-time data to improve production planning, leading to significant cost savings, reduced waste, and improved overall operational efficiency. This optimization extends to sustainability initiatives, as AI-driven control of resource planning can minimize the consumption of raw materials, water, and energy. AI can also analyze waste streams to improve recycling rates and reduce landfill waste, contributing to a more circular economy. Furthermore, it can enhance supply chain transparency and efficiency, optimizing transport routes and reducing fuel consumption, directly supporting corporate sustainability goals.

Agentic AI: The Next Frontier of Intelligent Assistance

Looking beyond the capabilities of current enterprise AI solutions, agentic AI represents a significant leap forward in automation and decision support. This emerging approach utilizes "digital workers" – sophisticated software agents that possess a remarkable degree of autonomy. These agents can perceive their environment by analyzing data from various sources. They can plan strategies to achieve specific goals. And, in a future vision, they will be able to execute these plans by taking action and seamlessly collaborate with other AI-agents. Furthermore, agentic AI systems are designed to continuously learn and adapt based on experience, making them ideally suited for the dynamic and complex environment paper industry.

Illustrative Future Use Cases for AI Co-Pilots in the Paper Industry

Imagine a mill manager in Canada needing to quickly assess inventory levels. They could simply ask their AI co-pilot, "What's the current inventory of 80gsm coated paper in the Vancouver warehouse?" and receive an immediate, accurate answer. Or a procurement team could utilize an AI co-pilot to automatically generate purchase orders for spare parts, taking into account delivery dates and prices from different suppliers.

Quality control is another area ripe for AI-driven improvement. An AI co-pilot could analyze real-time sensor data from paper machines, identify potential quality issues, and predict the necessary maintenance work and take the availability of the necessary spare parts into account when planning the maintenance order. AI Co-pilots can then also help maintenance personnel decide whether it's more profitable to keep the system running, bring maintenance forward, or postpone certain orders until after the maintenance is completed.

Production planning could be revolutionized by AI co-pilots that analyze demand forecasts, raw material availability, and machine capacity to suggest the most efficient paper and sequence. Predictive maintenance, powered by AI analyzing historical data, could anticipate equipment failures and prevent costly downtime.

Furthermore, these AI co-pilots could foster seamless collaboration between departments, such as sales, production, and logistics, across geographical boundaries. A sales representative in the United States could, in the future, use an AI co-pilot to expedite a customer order, coordinating with production facilities in Canada and logistics providers to ensure timely delivery.

It's crucial to understand the distinction between current capabilities and future potential. While many enterprise AI solutions offer robust automation and data analysis features today, the full vision of autonomous agentic AI, as illustrated by some of the co-pilot use cases, is still under development. However, the trajectory is clear: AI is rapidly evolving to become a more proactive and capable partner for human workers.

Embracing the AI-Powered Future: A Strategic Imperative

Enterprise-grade AI, and the future promise of agentic AI-powered co-pilots, offer the North American paper industry a powerful toolkit to address the pressing challenges of the skills gap, improve operational efficiency, boost sustainability, and gain a significant competitive advantage. By strategically embracing these technologies, paper manufacturers across Canada, the United States, and Mexico can unlock new levels of productivity, reduce costs, foster a more sustainable future, and secure their position as leaders in a rapidly evolving global market. As AI continues its rapid advancement, the opportunities for innovation and growth in the North American paper industry are truly transformative.

Stay informed on the latest AI trends, best practices and innovations from SAP by signing up for our newsletter.  

Artificial Intelligence (AI) is not only a buzzword in the 4^th industrial era. It is a real resource that is being adopted for a multitude of applications. In industry it is in its infancy, but as it matures adoption will grow.

One example of the growth of AI is in the power industry. The IEEE Power & Energy magazine for the November/December issue is entitled “Artificial Intelligence and the Power Grid”.  This issue is filled with examples of AI applications from around the world for monitoring the grid, forecasting load and generation, managing distribution, predictive maintenance, diagnosing faults, and optimization/control. It is impressive to see how the power industry has advanced in the implementation of AI for these purposes.

These advances did not come easily. The development of modeling algorithms has gone through many iterations to get to the point of efficient execution and sufficient accuracy. One of the challenges has been overfitting. Neural networks are a common AI modeling approach due to its universal application and relative ease of implementation; you don’t need to design a model on first principles. That is why overfitting can be a challenge. Overfitting is a problem in neural networks and several other modeling techniques where the algorithm learns the noise and outliers in the data instead of realistic relationships. In an article describing an implementation of convolutional neural networks to forecast electricity consumption in China from transformer data, the authors state “Transformer-based models have a substantial demand on the volume of the input data; for relatively small-scale datasets, they may introduce unnecessary complexity, leading to overfitting.”

This can be addressed by combining first principles into an AI model. In an article describing an implementation in Denmark, the authors introduce the application of physics informed neural networks which “learn directly from the equations describing the physical system.” In addition to improving model accuracy by mitigating the overfitting problem while using measurement data, the models are much more computationally efficient.  

Even when you have a sufficiently accurate model with efficient execution, it can be difficult to interpret the results. AI models tend to be black boxes; if the model is derived exclusively from data with no understanding of what the model looks like, it is a black box. In an article about virtual power plants in China, the authors state “AI interpretability is one of the key requirements for industrial AI applications. For example, when using AI to assist users in energy management at the edge, it is vital to explain whether the optimization solution offered by the model adheres to the production rules of the industrial assembly line and matches the user’s power usage habits. Only when AI solutions can provide a clear explanation of the reasoning behind their results can a credible and persuasive energy management plan be established.” Another article about South Korean AI applications states “Complex neural networks are often considered as hidden mechanisms, making it challenging to interpret their internal workings.”

A technique for opening the AI black box is called sensitivity analysis. If you hold all inputs of a model fixed and only move one from its lowest to highest range, a plot can be generated showing the relationship between that input and the model output. This can be very helpful for identifying overfitting in a model, and can help operations move their process toward an optimal condition. However, if there are correlated inputs this analysis can be misleading.

Another challenge we see in a lot of industrial AI applications is the difficulty in obtaining a good dataset. The editor’s column in this issue describes the history of AI in industry including “the difficulty of creating datasets suitable for the learning problems.” As we know, our data has noise, outliers, and process dynamics. If this data is not pre-processed properly, your model will learn nonsense. It is also challenging for a dataset to reflect sufficient process excitation to capture significant change. Some processes are single setpoint dominant, meaning that when they are operating they run the same way every day. Even processes that have grade changes may not have large step changes that produce sufficient excitation in the data to learn from. We don’t operate our processes with the intention of creating statistically significant datasets.

That is why I developed a dataset generator, which was the subject of a paper I presented at TappiCon 2023. The purpose of the dataset generator is to address scenarios such as:

• A student or young engineer that wants to learn data analytics
• An instructor that wants to teach students and young engineers data analytics
• A vendor that wants to develop a solution using data analytics, test the solution, and demonstrate it
• An engineer for a manufacturer that wants to develop a methodology for performing data analytics using various techniques

If the use case is to perform data analytics or predictive modeling for an industrial process, of course the actual data needs to be used. Actual process data cannot be replaced with artificially generated data. However, in the use cases considered, actual process data is not required. The goals of this project were as followed:

• Provide a general-purpose tool that can be customized and configured to produce a dataset representative of an actual industrial process
• Reduce the time to generate that dataset from the months required for actual data to hours for simulated data
• Provide the tool as open source, which can be obtained, customized, and improved at no cost.

With the dataset generator, I can develop AI applications and demonstrate them without asking anyone for their data.  

An example of this is an application I have developed in Seeq, which is a leading data analytics platform. The application automatically performs the dynamic pre-processing so that data from far back in the process (such as refiners) is aligned with the final product (at the reel). Models can be built very quickly and compared with different techniques, and the sensitivity plots are automatically generated so that overfitting and first-principle relationships can be analyzed. The dataset generator made it possible for me to build, test, and demonstrate this Seeq application with a sufficient dataset at low cost.

The last step is the quality and availability of data. We are often dependent on lab results in our datasets, and some lab tests have inherent variability that can exceed that of process instrumentation. In the paper industry we often lack measurements that are significant in the development of a model. An example is freeness. Lab samples once a shift will not capture the swings that explain production problems on a machine, and freeness testing in a lab has significant variability even when carefully following a standard procedure. There are also measurements, such as crill, that can have a huge impact on modeling paper strength but are not found in most mills. Today there are fiber analyzers that can provide very important information about furnish, making AI models more feasible.

The power industry is impressive in their adoption of AI. The paper industry can do the same with the right tools and expertise.

Here at Paperialo Publications we are using more and more AI every day (but will never use it to write our orginal opinion and analysis pieces). It is cheap and easy to learn. I think it will be as big a change in the way we do business as the Internet was along about 1994.

Here is what scares me--loosing AI wihout trained people behind it. The day will no doubt come when this won't be necessary, but it is not here yet.

Yes, some low level jobs can be eliminated, but for the bigger, consequential decisions, we are not there yet. When will that day be? I don't know and I don't think anyone will. And likely it will come in less critical areas first.

Do you want AI operating an airplane? Likely not. An automobile? We are on the cusp of that. A paper machine? Certain tasks now, others down the road.

Be assured the AI cat is out of the bag and it is not going back.

The way we receive goods has come a long way. Not long ago, deliveries relied heavily on manual processes, rigid schedules, and basic tracking methods that offered little transparency. In today’s fast-paced, e-commerce-driven world, speed, accuracy, and reliability are no longer just preferences—they’re expectations. Now, the Internet of Things is revolutionizing logistics and setting new standards for moving goods from point A to point B.

The Internet of Things (IoT) is still shaping the world around us and has been integrated into our everyday lives. Recent IoT developments are advancing swiftly, significantly enhancing various aspects of daily life.

The national absence rate in the U.S. was 3.1% in 2023, showing a downward trend from 3.6% in 2022. However, absenteeism still costs businesses $225.8 billion annually, averaging $1,685 per employee. Some industries face higher absence rates, with healthcare support occupations reaching 4.7%. Meanwhile, the agricultural sector has the lowest rate at 1.8%.

Educators and engineering firms looking for training on Industry 4.0 have a new resource: a six course curriculum for smart manufacturing created by the American Society of Mechanical Engineers (ASME) and Autodesk, Inc. The two organizations collaborated through 2023 and 2024 to compile interviews, analyze data and come up with real-world examples for the set of six free online courses. The lessons cover evolving engineering skills, including Artificial Intelligence (AI) and robotics, design for sustainability, Industry 4.0, data skills and business and digital literacy.