CHICAGO — Artificial intelligence.
Five or six years ago, AI that could almost think and create was something only encountered in science fiction.
How quickly technology advances. In the past few years, generative AI options have multiplied in almost every area of daily life and in business operations. That includes industrial and institutional laundries.
What started with generating content for business-related tasks (think communication, reports, documents) has moved onto the wash aisle. Insiders have even posited that some laundries could be operated by AI without any, or at least very limited, human involvement.
It begs the question, “Where is AI heading in the laundry industry?”
American Laundry News sought input from three industry veterans who are familiar with AI, what it’s doing for the industry today, and what it might provide in the future.
David Bernstein is the founder of Propeller Solutions Group in Livingston, Texas. Rodrigo Patron is director of operations for Lace House Linen in Petaluma, California. David Griggs — one of this magazine’s columnists at large — is director of operations development with Superior Linen Service in Tulsa, Oklahoma.
Where is artificial intelligence going in the industrial/institutional laundry space? What are some near-term enhancements that are coming?
PATRON: AI in the industrial laundry industry is clearly moving toward smarter automation and better operational visibility. Soon, we will probably see more systems that can automatically analyze production data, identify inefficiencies, predict machine failures, and optimize wash programs in real time.
We will also likely see stronger integration between production, maintenance, utilities and logistics systems. Even at Lace House, where we currently use AI more on the administrative side, it is easy to see how it could eventually become a much more useful operational tool as the technology becomes more affordable and accessible for mid-sized laundries.
GRIGGS: We will be running our automatic lint-blowdown fans throughout the day. Our floor-cleaning robot will ensure this lint is cleaned up from the floor just as soon as it lands. (I love watching these robots work at our local convenience store.)
AI is in our office, taking calls, making order adjustments, and invoicing.
I have seen an on-premises laundry (OPL) that sends the linen for each floor via an automatic rail system. There is no reason that filling the cart can’t be done with a robot.
BERNSTEIN: The clearest near-term enhancement is in maintenance management. Computerized maintenance-management systems (CMMS) powered by AI are evolving rapidly, and these new capabilities will soon find their way into laundries.
The leading CMMS platforms ingest real-time data feeds, establish baselines of normal equipment behavior, and flag potential anomalies before failures happen. They can estimate remaining useful life on critical components and trigger work orders automatically. Some can generate preventive-maintenance procedures directly from equipment manuals using generative AI and provide technicians with natural language access to asset history and troubleshooting guidance.
Intelligent parts-inventory management, driven by failure pattern analysis and supplier lead times rather than static minimum and maximum rules, is another near-term capability with direct application to laundry operations.
Beyond maintenance, I expect continued improvement in quality control systems, more sophisticated plant-scheduling tools, and greater integration between customer management and route accounting systems and production floor operations. The operator who can take a rush order at noon and have the plant automatically reprioritize operations, without human intervention, is close to fruition.
When is AI going to “think” for laundries, run wash programs and improve upon variables without human interaction? What would that look like?
GRIGGS: We already can have our daily load requirements automatically updated as each item is shipped.
In our dark (almost no human intervention) plants, we will be practicing predictive maintenance versus the reactive maintenance we all seem to get bogged down with today. Air cylinders, belts and bearings all have a finite lifespan. Replacing these parts on a set hour reading will keep our plant in motion.
I am assuming that the automatically fed laundry machines will have fewer misfeeds and jams, so no more jam pullers. I can see the entire maintenance staff just being contracted from the machinery vendors.
BERNSTEIN: Narrow autonomy already exists. A chemical dosing controller that self-tunes its formulas across cycles without human intervention is, in a meaningful sense, already thinking for a portion of the operation. An RFID system that classifies and discards its own bad reads using a neural network is doing the same.
But true plant-level autonomy, where AI is simultaneously managing load sequencing, chemistry, labor deployment, equipment scheduling, and customer priorities without a human making the calls, is a different order of magnitude. My honest assessment is that we are still some distance in time from seeing that, even in the earliest of early-adopter facilities, and longer than that for the broader industry. The prerequisite is solving the data fragmentation problem I described previously.
There are plants where every piece of equipment, from soil-sort through to pack-out, comes from the same supplier, and in those operations, data fragmentation is largely a non-issue. But even there, full autonomy would almost certainly require retrofits to machinery controls, the addition of sensors, and other infrastructure investments before the plant could function without human decision-making in the loop.
The technology to build a plant-level AI operations system largely exists today. What does not yet exist in most operations is the clean, integrated, real-time data feeds that such a system would require to function reliably. Someday soon, the pieces will come together.
When that day arrives, the plant will look similar but function in a fundamentally different way. The supervisory role shifts from making decisions to monitoring the decisions the system is making and intervening when something falls outside the parameters the AI is equipped to handle. That is still a human job, just a different one.
PATRON: AI is already starting to “think” for laundries in limited ways by automatically adjusting formulas and recommending operational changes based on data. Over time, those systems will continue to improve and learn from plant performance.
I believe there will eventually be systems capable of managing many operational variables automatically, including load sizes, wash formulas, chemical dosing, and production flow. However, I do not believe laundries will become fully autonomous anytime soon.
Laundry plants are still very hands-on operations that require people to troubleshoot issues, maintain machinery, monitor quality, handle customer requests and make judgment calls that technology simply cannot fully replace.
Come back Thursday for the conclusion about the possibility of “dark” laundries, along with final thoughts from Bernstein, Patron, and Griggs.
Click HERE to read part 1 about the current uses of AI in laundries. Read part 2 HERE about more benefits of AI use in laundries and challenges to overcome.
Have a question or comment? E-mail our editor Matt Poe at [email protected].
