It has been roughly 13 years since Marc Andreessen called out that Software was Eating the World, citing how internet-based technology companies were disrupting traditional businesses. Fast-forward to 2024, and it seems like Artificial Intelligence is Eating Software now.
This seems especially true in manufacturing, where AI is quickly making its way into factory floors across the world.
In this blog post, we will examine Industrial AI — what it is and what it isn’t — and describe its potential to play a crucial role in shaping the future of manufacturing.
All of today’s tremendous hype and activity around Artificial Intelligence can be traced back to November 2022, when OpenAI launched ChatGPT-3.5. This launch marked a significant milestone in the adoption of AI technology, both among the consumer and business sectors. ChatGPT rapidly gained traction with the masses, reaching a million daily active users within five days and a hundred million in just two months. To put it in perspective, it took TikTok about nine months and Instagram two and a half years to reach a hundred million daily active users.
What was particularly noteworthy about ChatGPT’s rapid growth is how even industrial sectors, traditionally slow to adopt new technologies, began integrating generative AI into their software stack by early 2023. Since then, AI assistants and Copilots built on large language models have been embedded across most software companies in the industrial technology landscape. Additionally, a few small language models have also emerged to address some of the accuracy, latency, and energy consumption challenges.
Either way, the rise of these language models has cast a spotlight not just on Generative AI but on Artificial Intelligence overall. This attention, along with subsequent market activity and venture capital interest, has inadvertently led to some measure of AI washing, ultimately resulting in widespread ambiguity.
This begs the question, “What is Industrial AI?”
To put it simply, Industrial AI is the application of Artificial Intelligence techniques, such as machine learning, deep learning, natural language processing, search, computer vision, causality, and robotics, in combination with industrial subject matter expertise, such as manufacturing operations fundamentals, heuristics, industry standards, asset failure libraries, process know-how, workforce-related information, etc., to solve critical problems for the manufacturing industry.
As mentioned earlier, it is important to note that Industrial AI is not the same as Advanced Industrial Analytics (AIA), which encompasses a variety of other analytic techniques in addition to AI, namely first principles, physics, and statistical models across descriptive to prognostic levels.
A common misconception about this topic is that Descriptive and Diagnostic analytics involve first principles, physics-based models, and statistics, while predictive, prescriptive, and prognostic analytics call for machine learning and artificial intelligence. As most of us in the industry know, that is not true.
Industrial process engineers have been using rules and heuristics-based root cause analysis, what-if analysis, and physics-based simulation to get predictive and prescriptive insights for several years without the need for AI. On the other hand, unlike traditional AI, generative AI’s reach goes beyond just analytics; it transforms how all software functions.
For instance, generative AI has the potential to automate repetitive tasks, enhance information consumption and analysis, build dashboards and workflows, and personalize user experiences across various industrial applications, such as CRM, ERP, EQMS, MES, CFW, etc. By enabling us to communicate with software through natural conversational language, Generative AI pushes the boundaries of human-machine-software interaction one step further since we evolved from command line prompts to graphical user interfaces (GUI).
Generative AI is a unique phenomenon within AI since it is both an AI model in itself and an enabler to build, create, and improve on other AI models. For instance, it can be integrated with an optimization model to make it more human-understandable and also used to program such a model. Furthermore, Generative AI has the potential to eliminate the necessity to learn trivialities such as the syntax of multiple programming languages or the exact folder path of a particular file.
So, to summarize, Industrial AI is not the same as Advanced Industrial Analytics. Some parts of Industrial AI, namely the machine learning and deep learning models, will continue to coexist with some of the other first principles and physics-based models within Advanced Industrial Analytics. On the other hand, Generative AI will play a much bigger role as it enables intelligent Copilots that assist with not just running analytics but also how we interact with software in general.
There is no question that Industrial AI will play a dominant role in shaping the future of manufacturing. Given the recent developments in computing, the continued relevance of Moore’s law, and the ability of AI models to learn and adapt, Industrial AI is on a rapid path to improve efficiency, productivity, and creativity across manufacturing.
Maintenance, process analytics and control, productivity, sustainability, and supply chain optimization are some of the common areas where Industrial AI is already making an impact. Machine learning and deep learning models trained on operational data help achieve step-change benefits in key operational metrics like unplanned downtime, throughput, OEE, etc.
In comparison, Generative AI has the potential to disrupt processes across the entire value chain, including information retrieval, summarization, assisted learning, and assisted programming. On the people side, Generative AI could also play a big role in addressing the skilled labor shortage in manufacturing by empowering frontline workers and the frontline leaders who specialize in managing, supporting, and solving problems for the frontline workers.
Compared to other technology trends in the recent past, I am particularly bullish about Industrial AI and its ability to create value at scale for manufacturing for three reasons: 1) It is rooted in problem-solving, 2) It serves multiple user personas, and 3) It can be applied to both short-term quick wins and long-term business cases.
For manufacturing companies actively pursuing Industrial AI, here are a couple of recommendations to get started:
Lead with business case and user persona impact: One of the common reasons why most strategic initiatives fail to scale is that they are rooted in technology and do not solve business problems. Similarly, while Industrial AI has the potential to disrupt several aspects of manufacturing, it is critical to approach it by keeping C-suite objectives, existing strategic initiatives, user personas, and use cases in mind.
Good AI requires good data: As many manufacturers continue to fast-track AI initiatives, it is important to note that a robust Operational Architecture, built on a solid data model with high-quality and contextualized data, managed and governed by Industrial DataOps, is essential to getting the promised results from Industrial AI.
Industrial AI is here to stay, but so are the people: The manufacturing industry is not the first, nor will it be the last, to hear the AI-is-going-to-take-away-jobs Industrial AI’s ability to automate repetitive tasks and enhance human-machine-software interaction will most likely impact the manufacturing workforce to some extent.
More importantly, Industrial AI will improve the nature of most jobs and create a variety of new ones, exacerbating the current skilled worker shortage. To meet this increased demand for skilled labor, manufacturing companies will need to empower their workforce to upskill and cross-train themselves in the near future.
In other words, people will not lose their jobs to AI, but they probably will lose it to the next person who can work with AI.
The three pillars of IX—people, process and technology—are all vital to success, but today we discuss the pillar that is probably the most important...
20 years ago, businesses were facing a crisis with their IT systems; Y2K. Just as Y2K was identified as a critical challenge several years in...
In this webinar, LNS Research details how manufacturers today are leveraging the Internet of Things to better connect to customers and suppliers.