AI and Automation: Progress Starts with the Process

This article appeared in the October 2026 issue of MiMfg Magazine. Read the full issue and find past issues online.

Artificial Intelligence (AI) has the potential to transform the manufacturing industry, and indeed, many Michigan manufacturers are already seeing the benefits of adopting AI tools. But there are two important caveats: Sound processes and decisions must come first — and people who have a deep understanding of operations and workflows should form the foundation of your AI approach.

“The real value comes when you combine AI and automation with good processes,” says Josh Johnston, Lean Program Manager for the Michigan Manufacturing Technology Center (MMTC) in Plymouth. Johnston and MMTC have been at the forefront of applied AI adoption, training and solutions. An increasing number of companies are now turning to MMTC with a fundamental question: Where do we begin, and how can AI improve our business?

“Traditional manufacturing processes are still going to be the backbone of any successful business, but what AI and automation are doing is actually helping us identify and respond to variability, improve visibility and really just speed up how quickly we’re able to respond to problems off and on the floor,” Johnston says. 

Johnston cautions about the temptation to use AI to “fix” a broken process that needs to be addressed at a more foundational level.

“If your process is bad and broken, automation alone isn’t going to fix it. It’s just going to allow you to do that bad process faster,” he says. “If a business has a disorganized process, adding technology layers on top of it is just going to amplify the pain that they’ve already experienced.”

The converse of that, however, can be a significant value-add when a manufacturer has strong fundamentals in place.

“When they have solid processes, AI and automation can amplify that performance in a really meaningful way.”

Strong manufacturing processes can provide the foundation for successful AI implementation but equally important is defining AI’s role and using it to complement — not replace — human decision-making.

First Things First

Automation and AI solutions can help manufacturers in so many ways that choosing the right approach can be a significant hurdle. At Chelsea-based Hatch Stamping Company, AI solutions are being implemented across the company to create efficiencies and improved workflows. Taking a measured and thoughtful approach with lots of input from their team has been a key ingredient in their success so far, according to Suzanne Morrison, Director of Marketing for Hatch.

“When you embark on a new technology or implement anything that is new and innovative, the approach that we take at Hatch is to start with the end goal in mind,” says Morrison. “We think about what can make the company better, stronger and more innovative.”

Since the company’s founding in 1952, Hatch has been a leader in tool and die and metal stamping —  so, innovating and adapting to new technologies is nothing new to the longtime Michigan manufacturer. Knowing that everyone from executive leadership to operators on the production floor were likely using AI tools already, Hatch wanted to gather input from employees before deciding on what tools to focus on and where they would have the most benefit.

“We decided early on that we needed to talk to the employees and find out where they were on their AI journey,” Morrison says. “So, we met with all of the departments in small meetings and, as you can imagine, what we discovered is that some were using AI a lot already and others not so much.”

After engaging with their employees, Hatch decided that starting with the basics was the best approach with the goal of creating efficiencies in daily workflow and therefore giving more time back to the employee for critical thinking and problem solving.

“We focus on using tools that will make our teams more productive and give us opportunity for better critical thinking,” says Morrison.

The “crawl, walk, run” approach of starting with basic tools has been successful for Hatch and can be a good model for small to mid-size manufacturers who want to avoid making a substantial investment.

“We didn’t go out and buy some big, expensive tool and, for most small to mid-sized manufacturers, that’s not an option for them and it’s not necessarily the best option, in my opinion,” says Morrison. “Making a larger investment down the road, that may be warranted eventually, but not initially.”

AI Applications and You

Robots are getting smarter every minute — at the tasks they’re given. What they can still struggle with is the broader process context and causal relationships that experienced employees understand, Johnston adds.

Take the scenario of an AI-powered camera used for quality control inspections. The camera can detect that a product is scratched but by itself it may not know that the scratch occurred hours earlier on another machine due to exposed metal. That’s where human intelligence and problem-solving skills are essential, and AI can free up employees’ time to focus on those issues.

“Instead of spending those hours looking at widgets, let them go back into the process, identify those patterns of failure and resolve the problems that created the defects in the first place,” Johnston says.

At Hatch, AI efficiencies have been implemented across the company from human resources and training to a proprietary application that helps operators gather data from production equipment. These tools exist to help the human ingenuity that has helped drive Hatch’s success for over 70 years.

“We’re working on custom tools that specifically align with our processes, and help collect data for production that supports a higher ROI,” Morrison says. “The outcome that we’re hoping for is time —  it’s value-added time for our employees.”

Safeguarding Your Data

Despite all the benefits of AI, the ability to protect sensitive information remains an important concern for every industry as well as individual users. To protect data, businesses need clear governance around how employees can use AI, including policies for approved tools, sensitive data, human review and accountability, says Johnston.

“Without that kind of policy, manufacturers are running the risk of their employees using AI for tasks that could pose a risk to the company — like exposing confidential information such as proprietary designs or customer details or even personal identification numbers, addresses or phone numbers of their employees,” Johnston says. To show manufacturers “the art of the possible” when it comes to data security, MMTC has built an AI lab at their Plymouth location to demonstrate how an on-premise solution reduces exposure of sensitive company data by operating in an isolated environment not connected to the Internet.

Part of the implementation and training at Hatch focused on the importance of protecting sensitive data while using an open-source AI tool.

“Confidentiality is a priority when we’re working within departments like HR, but really we have sensitive data in every department,” says Morrison. “Even though (one of our employees) may purchase one of the AI tools available they need to be extremely careful about what information they share.”

Systems of Experts: Capturing Tribal Knowledge

Specifically in regard to training, leveraging AI does have some urgency.

AI reduces the need for workers to understand each nuanced step of technical or tedious processes, reducing burnout and increasing efficiency. At the same time, someone must understand those steps to identify, prevent and solve problems.

In some cases, that knowledge exists only in the minds of soon-to-be retirees. With Generation Xers and the youngest of the Baby Boomers reaching retirement age, manufacturing leaders are concerned about an exodus of employees who will take that foundational knowledge with them. To complicate matters, because automation has or will take over those technical steps, there might not be as much pressure to share that knowledge with others on a daily basis.

It’s therefore important to document that pertinent information before it walks out the door. Johnston says using self-service chatbots that contain answers to questions, known as Systems of Experts, can be especially helpful.

Expert systems date back to the 1960s but generative AI has made interacting with organizational knowledge far more conversational.

“Then, you would walk up to one of these terminals and go through these ‘yes-no, if-then’ statements to get answers,” Johnston says. “Now we can capture and organize an expert’s knowledge and make it conversationally accessible to people across the company.”

Training and Adapting the Workforce

AI tools can be user-friendly and adaptable which provides a big leg up for the training aspect. While the new workforce doesn’t need to be made up of programmers and developers, workers do need to be comfortable working alongside technology, Johnston adds.

“One of the big misconceptions is that AI is coming to steal all of our jobs,” Johnston says. “It’s not AI that’s going to steal your job, it’s somebody who knows how to leverage the technology that’s going to steal your job.”

If there is one skill that workers must develop, however, it’s data literacy.

“Data is the new oil in terms of value to companies because data is the underlying factor that directly affects what AI and advanced automation can do,” Johnston says. “We need to understand what good data looks like and how to ensure that it is able to be consumed by these automated or AI systems.”

Humans should always remain in the loop to ensure that information generated by AI is accurate, Johnston says, adding that companies that will benefit from AI the most are those that stay focused on practical applications and the value of the solutions.

“As long as companies keep that kind of focus, the future is going to look bright in AI and automation,” Johnston says.

Have a manufacturing story to tell? Share your insights at bit.ly/insidemfg.