You know what it's like to have a craving for something?

On a recent trip to London, I had a craving for a BLT (Bacon, Lettuce and Tomato sandwich). You can imagine my disappointment when I heard they were out of tomatoes. It made me think about possible root causes of why such a commodity would even be short on supply. What came to mind was the climate change that is impacting crops, soaring energy prices which cause UK greenhouses to shut down during winter, a shortage of labor which translates to no truck drivers, and all this topped with geo-political reasons that restrict movement of goods and divert them to other countries.

I understand that not having tomatoes for my sandwich sounds a little trivial, but for all businesses – from Amazon to Zara - a shortage of supplies or materials for products or packaging can be a question of survival. When faced with supply chain disruptions, a business doesn’t “need AI”, they need what AI can do.

Let me explain:

Amid all the hype and excitement around GenAI and ChatGPT, it is important to remember that AI is not new. Nor is the Intelligent Enterprise. Cooler heads must prevail. There is a fear of missing out (FOMO) mentality that threatens to overwhelm good business sense. Today’s businesses need to be agile to remain competitive in a constantly shifting landscape. ERP solutions can enable intelligent enterprises to face emerging challenges and to easily anticipate and address changes, leading to better productivity, visibility and results. AI can magnify this.

But how can companies leverage this valuable capability to improve ERP software, by simplifying and enhancing the current functionality and ultimately making the software easier to consume? And how exactly can businesses benefit from an AI-enriched ERP and begin their transformation into an intelligent enterprise?

In the case of my sandwich, I’m confident AI-enhanced ERP could have helped: by monitoring water levels to improve crop yield, or else enabling efficient delivery planning regarding transportation by identifying the available transportation routes and resources using real-time information.

We all know, resilient supply chains are the backbone of every successful business and economy. This is an area in which AI provides massive benefit: AI can predict, prepare and help companies appropriately react to today’s fast-evolving demands. Based on a product’s demand, characteristics or forecast, AI can propose means of warehouse organization, in addition to stock and replenishment processes optimization. AI-enriched applications can automatically calculate demand and order products to allow retailers or transporters to plan and optimize replenishment across their distribution centers. This not only leads to optimization of inventory levels and costs but also boosts product availability and margins.

By forecasting the risk of a late payment on an invoice and helping with customer prioritization an AI-enhanced ERP helps finance teams to quickly identify changes, reduce risks and makes it easier for them to control costs. Sales teams can spend more “quality time” with their customers, thanks to AI completing complex sales order documents automatically, enabling faster order processing and making manual order entries by sales reps obsolete. Procurement processes can be made more productive and intuitive and support within ERP software is made more responsive using the advantages of AI.

A balanced diet of AI innovation

We have been working on AI innovations for several years. Sometimes, amid the whole GenAI hype, one tends to forget that AI is not new. It’s simply called “more traditional AI” now. These innovations can be found everywhere. For example, the mapping of certain documents together using artificial intelligence, which reduces the need for tedious, repetitive and error-prone manual work. Or in manufacturing, they are used for visual inspection, where cameras are used to check the materials for potential quality issues at multiple points in the milling and manufacturing process.

I believe it’s about taking away a fear that’s sometimes out there. The world is constantly changing, business is adapting to continuous disruption and now technology is being revolutionized with AI. Throughout all this, having a trusted partner and confidence in your data enables you to move your business forward. Software providers already started putting AI and machine learning in their solutions years ago, but for a long time, customers have been hesitant. ChatGPT and generative AI have made AI feel more accessible, more tangible and many customers are now more open and keener to use AI capabilities.

It’s important for vendors to say: “It’s not about AI, it’s about what AI can do for your business”. Once a business knows where they’re heading or what differentiates them, what their customers are looking for, then they can ask, “how can AI help to achieve this?”. The right ERP provider will help their customers to improve how they run their business and that includes showing them where AI can fit into it. They take care of the complexity and journey with them in that partnership. Customers don’t need to be an AI expert, they just need to be an expert in their business.

ERP has been at the center of business for decades, enabling businesses to lower costs, improve quality and provide a stable platform for innovation. To leverage innovation, you have to have cloud scale. To get benefit from AI and become an intelligent enterprise you must be in control of data quality. You must have cloud-based systems, you can’t do all that stuff sitting in an on-premise data center. Standardized ways of working are key: industry best practice processes or out-of-the-box public cloud solutions for everything that doesn’t differentiate you.

To me, the future of AI is extremely exciting, and I believe AI will fundamentally change the way we work, change the way business functions, and will obviously change how software is being used in a company.

There is so much potential when it comes to injecting AI in ERP solutions. Though with all the current hype around this topic, it’s important to remember that why you’re in business, what differentiates you, and how you serve your customers is what really matters. So, to avoid making a meal of your business, ERP providers must be what their customers need them to be: the rational advisor and trusted partner in a fast-changing business world, without the side of FOMO.  

Throughout my career in automation, which began in 1984, I have worked on advanced control. Therefore, advanced control is not a new concept.

However, we need to define what advanced control is. TAPPI TIP 1103-04 (Industry 4.0 Lexicon) defines it as:



• Realtime, deterministic closed loop control

• Not discrete

• Has at least one of the following attributes:

o  Model

o  MIMO

o  Advanced Regulatory

o  PID replacement

o  Cascade

To elaborate, it needs to be a control loop that may have multiple inputs and outputs, is running in realtime in a deterministic schedule, is regulating floating point values such as flow rates and temperatures, and often employs dynamic predictive models. MIMO refers to Multiple Input Multiple Output, meaning a control matrix that solves for meeting multiple targets using multiple outputs. A Proportional Integral Derivative (PID) controller is not advanced control because they have been around for a long time and are ubiquitous, but enhanced implementations of PID or replacements like Fuzzy Logic are considered advanced control. When a PID is put in Cascade mode, whatever is on top of it driving the cascade setpoint is advanced.

Given this definition, it is obvious to veteran controls engineers that advanced control has been around well before the 4^th industrial era. However, in today’s era advanced control has new requirements capabilities, and challenges.

Large scale advanced control application, such as MIMO, is often hosted in a server machine at Level 2. At this level, you can have deterministic processing and connectivity through OPC to the base level data and PID loops at Level 1. There are many proponents of putting everything into the cloud at Level 4. This represents a fundamental and dangerous misunderstanding. You will not get deterministic processing or communication from a server farm connected over the public internet. It is possible to have steady state optimization in such a deployment, but most practical applications in a paper mill require dynamic models and determinism.

The control matrix and prediction models in today’s era can be bigger than in the past. As systems are more connected, the opportunity for multi unit optimization exists. Previous islands of automation can be connected as a unified automation system, and therefore the advanced control application can grow in scope. As just explained, the urge to push deterministic applications into non-deterministic domains is a danger. Having experienced engineers that understand the process and the practical limitations of implementation can keep you out of trouble.

The biggest buzzword in advanced control today is Artificial Intelligence (AI), which is often associated with Machine Learning (ML). Again we need to go to definitions:

AI

• Uses process data
• Uses mathematical techniques to turn data into information or action
• Creates a system that mimics human behavior
• It does not have to be adaptive

ML

• Meets the definition under the umbrella of AI
• Adaptive

To elaborate, artificial means it is a machine (not a human), and intelligence means it has knowledge. AI behaves with human intelligence but may be programmed, configured, and trained by a human. ML has the added ability of being able to train itself (learn) with the data it can access. We are in the infancy stage of ML automation in industry. Data is not completely clean; there is noise and outliers that could cause ML to learn exactly the wrong relationships. If a badly trained ML is automated, it can be dangerous. However, AI is an umbrella term that encompasses all the advanced control techniques in practice for decades. Therefore, we have arguably had AI as long as we have had advanced control. ML will evolve carefully until there is comfort in knowing that it can learn the right way.

As a veteran advanced control engineer in the pulp and paper industry, I see a lot of opportunity and tools to facilitate implementation that did not exist when I first started. However, the principles are the same. It is important not to get caught up in hype and buzzwords to make 4^th industrial era investments pay off.

Jan is bemoaning a lack of tomatoes in the lead article this month. This is representative of a problem I have seen elsewhere.

Since 2020, construction projects have been beset by lack of critical items. In another column I write in another one of our publications, I commented recently that I never saw "Force Majeure" in any place but contract language up until a couple of years ago. Now it has become the lingua franca.

I remember a little over forty years ago, I had joined a company and was receiving orientation in several departments. In the purchasing department they told me how they could now track late shipments. The purchasing manager showed me on his computer screen that it now showed late shipments. For a very brief moment, I thought they had some devine way to track materials. About three questions led me to understand that a human had to call the provider, ask the status of the shipment, and then manually enter it in to the computer.

I could do the same thing with a pencil and paper, it just wouldn't be printed on "green bar" (does anyone even know what "green bar" is any longer?).

However today, with today's computers, software, Internet, and AI, we can do that which was my brief flash four decades ago. We are able to know before it is a problem where the delivery problems are.

Now, you might say, "Hey, Jim, you just threw AI in there because it is the buzzword of the day." No, you are wrong, for if I build a database of suppliers' performance, AI will be able to alert me way ahead of time of potential delivery problems. Build the database out further, and AI will be able to offer alternative suppliers in time to not affect my needs.

This is powerful.

Manufacturing has undergone a seismic transformation in the last decade, driven by technology advancements that are reshaping the landscape. Pillars like Enterprise Resource Planning (ERP) and Customer Relationship Management (CRM) systems are now table stakes, but they only lay the groundwork for what's to come. Today’s manufacturers are on the edge of a new industrial revolution driven by the collaboration between humans and machines to increase the focus on resilience and sustainability.

In the labyrinth of technological innovation, the Internet of Things (IoT) stands out as a beacon of connectivity, weaving an intricate tapestry that spans industries and revolutionizes daily life. As a seasoned journalist entrenched in the IT sector, my objective is to unravel the dynamic landscape of IoT trends, all while maintaining the subtlety of my professional background.

In a time defined by the Fourth Industrial Revolution (Industry 4.0), a period transforming industries through advanced technologies, the manufacturing landscape is evolving beyond something we recognize as familiar. With the potential to automate processes, enhance data visibility, optimize operations and redefine value chains, technology is revolutionizing industries globally.

Convergence is the integration of operational technology (OT) and information technology (IT). It is primarily used to solve key IoT challenges. Let’s take a look at its primary roles in the IoT value chain.