TEXTILES.ORG
Most companies likely would welcome an employee with a Ph.D. who works for $30 a month or a tool that makes employees more efficient.
That’s exactly what artificial intelligence (AI)—particularly generative AI—aims to do for businesses of all sizes. The difference between generative AI and other forms of AI is that traditional AI relies on algorithms and data to perform a specific task, while generative AI creates new content.
If business owners are wondering whether generative AI can assist their companies, Nelson Valderrama, founder and CEO of Intuilize in Grapevine, Texas, says yes. “I think this is for you because this is an opportunity to hire an assistant that has a Ph.D. degree in terms of knowledge,” says Valderrama.
AI is not cost-prohibitive for most companies, according to Dan Hughes, cofounder of Claritee.AI in Doylestown, Pa. Many AI subscriptions cost $20 to $30 a month. “Most technology, when it’s introduced, is not within reach of small- to mid-sized businesses,” says Hughes. “AI is completely different. It’s very much in reach.”
Eliminating time sucks
There are practical ways generative AI tools can help businesses enhance their workflows. For example, when processing invoices, emails with invoice PDFs can be saved to a drive and AI tools can integrate into the process.
“What’s happening is people are putting these AI models now to work inside of workflow tools like Zapier to analyze an invoice, extract the data off the invoice and then take action with that,” says Hughes.
Valderrama suggests starting with accounts payable before using generative AI for invoices because invoices affect both customers and finances. “Many major ERPs [Enterprise Resource Planning systems] already offer ready-to-use automation tools for this purpose,’” he says.
Manual data entry can lead to errors, but generative AI minimizes incorrect work by automating processes and flagging potential issues. For example, if an item price seems unusually high compared to past patterns, the system might notify you with a question: “The price for custom valances is 30% higher than usual. Is this correct?”
Creating training materials
AI is also proficient in writing instructions. Hughes demonstrated this capability by downloading a diagram of a wire harness in ChatGPT. After entering a prompt to create assembly instructions, AI produced results within seconds.
“These AI models have a vision capability—that is, it understands what it’s looking at,” says Hughes. “This is the new version of AI because it is reading and interpreting everything going on here. It’s not just looking at objects or keywords. It’s understanding what this is.”
One large time commitment for businesses is creating standard operating procedures (SOPs). However, Hughes says companies can record or take photos of an employee doing a particular task and transcribe the images to text. Ask AI to write a procedure about this specific assembly and an SOP is generated.
“People spend a lot of time writing and documenting SOPs, and all that can be fully automated,” says Hughes.
Digging into the data
Web scraping, or extracting information from websites, is an important AI function for businesses, according to Hughes. Many of the AI tools that perform this action can reside on browsers as plug-ins. “They’re impressive,” Hughes says. “You can pull data down, like pricing information.”
To demonstrate the power of generative AI, Valderrama asked Claude AI a series of questions about the textile fabrication industry by typing prompts for Claude to answer. One prompt asked Claude to analyze the cost optimization for a small shade fabrication shop in California that spends $18,500 a month on operations. Claude presented four ways to reduce costs and a graphic that displayed the breakdown in savings—all within a few seconds.
Getting up to speed
When it comes to learning and becoming proficient in recent technologies, there is no substitute for firsthand experience.
“How can individuals augment themselves to be more productive?” asks Hughes. “That’s the first step. Even before getting to invoicing and workflow automations, we always direct people to be more productive using the general-purpose chatbots. Just that alone is a giant step forward.”
Many robust AI tools are available, including ChatGPT, Claude and Gemini. “All of them are impressive with their basic capabilities,” says Hughes. “That’s message number one … leveraging those tools to increase your own personal productivity.”
Hughes points out that Microsoft Office and Google integrate AI into their products, including Gmail, Google Sheets and Google Docs. “And once you get your basic productivity—like your chatbot, your Office product tools—now you can start to look at those individual products to help enhance different parts of the business,” says Hughes.
Although learning and becoming proficient in new technologies can be daunting, Valderrama suggests focusing on the benefits AI can bring. “Just relax and really try to see what the possibilities or options or solutions are that it can provide,” he says.
Caveats about private data
Despite generative AI’s advantages, some concerns exist. One is the possibility that a company’s data could be compromised on a generative AI platform. Valderrama urges businesses to use caution about the type of information they enter into an AI tool. For example, companies should avoid putting their profits and losses in AI or entering Social Security numbers. “Anything that is extremely sensitive, don’t put it there,” says Valderrama.
Also, companies should consider paying more to protect their data when using generative AI. “When you’re using any generative tool, it’s extremely important that you get the proper licensing. Don’t use a free license for business. Use a paid one,” says Valderrama.
AI platforms offer different tiers—from free to paid—that treat data differently. Hughes cites an example with ChatGPT: The free version uses data to improve models. “They don’t actually take your data and put it in someone else’s chatbot, but the data is still being used to train the model,” he says.
If you select a higher subscription, like the Teams and Enterprise levels, ChatGPT will not use the data or track analysis that is conducted. “Then you’ll have a higher level of security,”
says Hughes.
Need for humans
The common concern people express about generative AI is that it will replace workers. Both Valderrama and Hughes say that certain human skills, such as creativity and judgment, are still essential—and oversight by humans remains necessary.
“The importance of using any of these things is having a human being that has the knowledge to validate this output,” says Valderrama. “Even with consistent use, [generative AI] typically provides responses that are accurate or relevant about 60% to 80% of the time. There will always be a margin—around 20% to 40%—where [it] falls short due to its limitations.”
This illustrates why human oversight is essential; generative AI results need to be reviewed by someone knowledgeable about the subject to verify everything is accurate.
For instance, when Valderrama asked Claude AI a question about regulatory requirements in the textile industry, it gave a response mentioning PFAS and provided examples. A human reviewer would need to verify the sources to confirm the results are accurate, up-to-date and relevant.
“This process is similar to supervising an intern; you review their work until you’re confident in its quality,” says Valderrama.
“It’s making us more productive, and it’s automating some of our workflow, but we still have to be in the loop,” adds Hughes.
Alan Pierce is a freelance writer in Burnsville, Minn., with a background in journalism as a reporter and editor.