Ehmke Manufacturing

Ehmke Manufacturing has been a Philadelphia institution since 1929. The company was founded by Howard Ehmke, a professional baseball pitcher for the Philadelphia Athletics, who opened a small cut and sew shop next to Connie Mack Stadium upon his retirement and made the first baseball infield covers. For decades, the company turned out awnings, canopies, and other simple sewn goods.

That changed in 1992, when Bob Rosania and his college business partner, both mechanical engineers, bought the company and began driving it toward more technically sophisticated work. Their watershed moment came in 2002, when Boeing nominated and awarded Ehmke its Global Supplier of the Year out of 14,000 companies. Today Ehmke has grown from 18 employees when Bob joined to roughly 150, with a product line that ranges from individual first aid kits to a recent award-winning multi-layer proprietary cover that protects C-130 aircraft from hail damage.

That diversity is the company’s signature. As Bob put it, “When you see the government shutting down or funding issues out of Washington, how we’ve been able not only survive, but thrive, is because of that diversity of product that we offer.”

Ehmke and DVIRC have been working together since 2006, when Boeing’s edict to implement a Lean manufacturing plan first brought the two organizations together. Bob has served on DVIRC’s board for more than a decade.

When the conversation around generative AI began heating up, Bob saw a familiar pattern. “The industrial textile industry is busting out of the 1990s,” he said. “We’re always so far behind on adopting technology, embracing technology, embracing robotics, whatever it might be.”

There was no acute crisis driving Ehmke to act. The company had navigated COVID and government shutdowns when many in its industry had not. The underpinnings were strong. But Bob and his team did not want to wait until AI became a problem to be solved. They wanted to understand the technology on their own terms and identify where it could create real value across the business.

“We thought it was important to gain an understanding of what it is, and then get our employees spun up on how we can use the AI as just another tool to set us apart, to distinguish us from others,” Bob said. “Let’s be some of the early adopters of this and use it to our advantage.”

The goal was practical: separate hype from reality, give a cross-functional team a working understanding of what GenAI could actually do, and develop a prioritized list of use cases the company could attack.

Ehmke engaged DVIRC for a hands-on GenAI training program led by Dan Hughes, Co-Founder of ClariteeAI and DVIRC’s GenAI instructor. 10 Ehmke team members participated, drawn from sales, finance, operations, quality, and senior leadership.

Bob had first encountered Dan Hughes at an industry Outlook conference and was impressed enough to refer him into DVIRC’s network. “We were pretty impressed by the way he approached things. Very practical, very right-sized for small and mid-sized manufacturers.”

The two sessions blended foundational GenAI concepts with workshop-style exercises built on real Ehmke information. Participants worked through prompt engineering, content generation, and the differences between models, learning that ChatGPT served as a baseline while Gemini tended toward more creative responses, and that none of them were 100 percent accurate without human review and refinement.

What set the engagement apart, in Bob’s view, was how specific it got. “It wasn’t just general topics that were brought up,” he said. “Literally taking Ehmke specific information and coming up with an Ehmke specific solution.”

Dan and team also helped the team draw a critical distinction between true AI use cases and what was really just data management, then walked the group through a structured process to identify and risk-assess specific opportunities inside Ehmke’s operations.

By the end of the program, Ehmke had developed a working inventory of 22 prioritized use cases. Examples include inventory management and vendor performance, shop floor scheduling that currently runs through interactive spreadsheets, sharper and more granular customer presentation responses, and a longer-term vision Bob called “transformational”: consolidating vendor performance, inventory, and capacity information into a single central database that AI tools could draw from.

The pacing also landed. Bob said, ” Dan and team did a nice job of dialing in at the level of expertise or lack thereof of our group. The novice users versus the more sophisticated users all got something out of it.”

Bob’s continuous improvement team is now working through the 22 use cases, conducting risk assessments and selecting which to move forward with first. “I see people more engaged in that than prior to the training,” Bob said. Team members are more comfortable using ChatGPT to drill down on a presentation, refining prompts, and treating the tools as a two-way interchange rather than just a place to look up information. They have a better grasp of model differences, of where the gaps are, and of why human interaction is still essential to refining any AI output.

For Bob, the broader takeaway for other manufacturing leaders is unambiguous. “Run to the light. Don’t run away from it. Whether you choose to incorporate it much into your business or not, you should fully understand it. AI is not going to replace the workers. It’s the workers who utilize AI most effectively are the ones that are going to replace you. So get your workforce coached up on what AI is about and how it can help them.”

His verdict on the engagement itself was equally direct.

“It was one of our best training sessions we’ve had with the DVIRC. It’s the right-size service at the right-size price.”

Bob Rosania, Ehmke Manufacturing

Ready to build AI fluency on your team? Get in touch with DVIRC to learn how a GenAI workshop can help your organization identify high-impact use cases and put practical AI tools to work.