.png)
1 week ago
7
Industrial AI has some amazing claims and lackluster performance. If we use AI as a plot device in manufacturing’s most well known novel – Goldratt’s The Goal – it helps us see where the hype cycle is too far gone, and where modern industry can expect to get payback from this new tool. The best way to look at AI is as a ‘super-excel’ that enables better distributed decision making throughout the organization, especially in specific sandboxes that are data rich that are not covered by current enterprise software systems.
Goldratt’s contempt for robots – what was then the ‘hot’ technology at the time he published The Goal in 1984 is palpable – we can feel him sneer every time the NCX-10 is mentioned. It’s a great literary device as the protagonist, Alex Rogo, realizes that upper management’s obsession with this one production step won’t improve his team’s performance. If anything, by spending time with the robot, his team is distracted from working on the major issues the business faces, and it prevents them from recognizing the system’s true constraints.
What if 1984 Goldratt had instead written science fiction and brought today’s AI back in time? Here are four ways I could see AI impacting the plotline based on what we’ve learned implementing AI solutions for a small manufacturing business.
1/ Like the NCX-10 Robot, AI Would be Useless.
The robot, the NCX-10, is introduced in the first chapter and is used to introduce constraints and other concepts in later chapters. In Chapter 17, we learn that the NCX-10 has a 25 unit max batch size, so other, previously retired production units, are brought back on line. When Rogo helps his team adopt Jonah’s approach to throughput accounting, they realize that the NCX-10 should simply be ignored.
Unfortunately, that’s also been our experience with several AI product claims. ChatGPT claims to have spreadsheet compatibility – that hasn’t worked for at least four weeks. We’ve trialed a number of different products and the claims have not matched reality. Like the robot, it’s not clear what an AI product really does – it’s a bunch of marketing hype with potential, but it isn’t necessarily the constraint within the system.
In the end, our view now in 2Q 2025 is that AI is a superior form of spreadsheet, or a way to introduce low-code solutions to a bigger portion of our workforce. A team member who is okay with coding can become very good. An adventuresome colleague who knows excel can dive deeper into python and automate more of their workflows. Activities that are repetitive are more easily automated, but AI is by no means a panacea. Rogo would likely have found the same.
2/ Could AI Replace Jonah? Absolutely Not.
Goldratt teaches us through the Socratic method. Alex Rogo derives his name from the latin verb ‘rogare’ which means to ask or to question, and is a clear stand in for the reader. Rogo’s mentor is Jonah, a transparent version of Goldratt himself.
Could a version of The Goal exist where the AI plays the role of Jonah? Unfortunately, this would be pure science fiction. No current AI model is capable of consistently answering these questions accurately. Models often lie. Models often make egregious factual errors. Even when confined to a single focused data set – which is surely what the Unico division of Uniware would have done.
3/ Was AI Ralph Nakamura’s Secret Weapon?
As a plot device, Rogo never wants for data. At every point when he faces a problem, Rogo never has to wait months to get an attempt at cost accounting or see recent production numbers. His jack-of-all trades from his team is Ralph Nakamura, who is always ready with the right information whenever he needs it.
Realistically, this is the best use of AI. If anything, Ralph’s solutions are always a little too convenient as a plot device in 1984 – maybe The Goal really is science fiction. Ralph should be running training sessions for the other team members and empowering them on how to use the tool. After all, MS Excel would not be introduced until September 1985 – after the book was first published.
4/ Improved Product Quality, a Happier Bucky Burnside and a Better ERP/MRP?
The book starts with a big customer, Bucky Burnside (not Bucky Barnes, Marvel’s Winter Solider) furious that his part 41427 which has experienced quality issues. This creates the typical ‘hot lot’ mentality at the plant, as Rogo and his crew work to locate, identify and produce the part. This is a good use of AI – in the same way that modern ERPs/MRPs are the most likely businesses to get disrupted by AI.
As a small domestic manufacturer of equipment for heavy industry, this is where we’ve seen AI be most useful. We are very systems lite. This is common in manufacturers of all sizes and scale – big manufacturers have sites that are poorly served by aged and brittle systems. Smaller companies have lightweight or non-existent systems that beg for more capabilities.
We’ve been successful in training our team to use Python and other software systems to automate reports and business processes that were previously done manually. For a small business with an office team of eight, we’ve identified over sixty hours in saved time per week created through AI automation. That’s a near 20% boost in productivity.
Conclusion: A Better Tool, Not a Magic Tool
Scenarios 3 and 4 are in line with our own experience testing out multiple AI platforms over the past two quarters. If we use it as a better version of excel, a better version of a CRM, and confine the use of the AI tool to data rich sandboxes, then the results are impressive. We can enable every team member to be their own Ralph Nakamura, and we can enable our organization to respond to every customer with the urgency that Rogo’s UniCo Great Barrington plant responds to Bucky Burnside.
AI can’t yet come close to replicating the 1984 productivity expectations of robotics or of Rogo’s Socratic mentor, Jonah. Just because the hype does not match reality in those areas doesn’t mean we should wait to implement these tools where possible.
The impact of AI will be reduced hiring of mid-level clerks and fewer outsourced AI jobs. For those coming in to the manufacturing industry, focus on learning how to work with the tools to automate work flows and make data-driven decisions.
