CHICAGO — Make no mistake, artificial intelligence—AI—is weaving its way into industries large and small, including the humble laundromat.
From optimizing maintenance schedules to predicting customer demand and even tailoring marketing strategies, AI can offer laundromat owners some surprising tools to streamline operations, boost profitability, and enhance the customer experience.
Even if it’s only put to use assisting with everyday administrative or managerial duties, AI might just be the key to taking a laundry services business to the next level.
GETTING TO KNOW AI
Before getting into its potential, it’s important to get a sense of what AI is and how it works.
For a small business, AI refers to computer systems that utilize algorithms and data analysis to automate repetitive tasks, improve decision-making, personalize customer experiences and gain insights from data.
Matthew Krieger, president of Cober Inc. and a small-business mentor for SCORE, described AI in greater detail during his webinar titled “Leverage ChatGPT and AI Tools in Your Business to Increase Productivity.”
“Generative AI is AI that can be communicated with and also communicate with you in a humanlike way,” he says. “The generative part is that when you ask the AI a question, it can actually create new outputs based on that question. AI is nondeterministic. That is, the results are being created in real time. So if you ask the same question twice, you may actually get two different but similar answers.”
An AI model is a computer program that is trained on very large data sets—including knowledge, human language, images, audio and video—to make decisions or to recognize patterns.
And a large language model—or LLM—is a type of AI model that specifically focuses on understanding natural language. ChatGPT, which runs on a set of language models from OpenAI, attracted many users shortly after its release and many competing models have since been released. Some belong to companies like Google and Microsoft, while others are open source (the code is freely made available for modification).
AI tools rely on user prompts—the questions, statements or commands entered to generate a response or action—that may or may not include supplied data. The better the detail and context, along with the quality and quantity of any data supplied, the better and more complete the response.
In the business world, there is what’s called structured data and unstructured data, Krieger says.
“Structured data is data that you’d find in a spreadsheet or in a QuickBook system. That is, there is a data definition, there’s a row, there’s a column, some data is text, some data is numbers, and the system understands the type of data that it is,” he explains. “Structured data can be thought of in Excel or Google Sheets when you want to sort on columns or sum of data.
“Unstructured data is data that lacks structure: handwritten notes, logs that you keep, a journal, a (Microsoft) Word document. All of these things are considered unstructured because it’s just freeform data. There’s no definition to that data. Well, that can present a problem because a lot of our data is unstructured. But we still want to ask questions of that data and get intelligence from it.”
During his webinar, Krieger demonstrated some everyday tasks that are possible using a few different AI-driven tools for both desktop and mobile. Some of them he used are paid but he says most of his examples can be done using free versions.
“All of them … are actionable by you,” he says of the demos. “Some of them may be practical to you but all of them should really serve to help you see what’s possible for your business.”
He’s been keeping a food log because he wants to start working with a nutritionist, Krieger says, but it’s 20 pages of dates, times and foods consumed. He wants a quick summary, so he uploads the document to an AI assistant called Claude.
“Tell me the major food groups I eat in descending order of consumption,” Krieger types into the prompt area and hits return.
Within only a few seconds, the cursor comes to life and produces a detailed list with proteins, vegetables and so forth. “What do I eat most?” he types. Another moment passes and Claude responds with a numbered list of food items, with avocado toast at the top. “What do I eat least?” It’s red meat followed by dairy, according to Claude. “How late do I typically eat dinner?” Most common time is 7 p.m.
Other demos generated sample data, analyzed fictitious financial statements, created visualizations from data, took inventory, and even located items on crowded store shelves.
In Tuesday’s conclusion: A list of administrative, managerial and employer/HR tasks that AI could do for you
