From Experimentation to Execution: AI's Real Manufacturing Impact

“There’s a story behind why he made certain decisions,” Navin explained. “That conversation is probably more important than the specific setup itself.”

The impact of that effort is already tangible. Navin described a recent case where a customer reported an unusual machine issue on equipment that had been in the field for years. By searching archived service reports, the team quickly identified a similar issue from a decade earlier.

“What would have taken hours, we found in about 30 seconds,” he said.

For Saad, that type of application represents one of AI’s most accessible entry points. “You can capture meetings, transcribe them, summarize them, and make that knowledge searchable,” he said. “That becomes incredibly valuable over time.”

Internal efficiency is the starting point

While AI’s long-term potential includes advanced automation and machine intelligence, most organizations are beginning with internal efficiency gains.

At ProMach, Cartia said the company is focusing on areas such as engineering, sales, and productivity. “We’re starting to see value in driving internal efficiencies,” he said.

Engineering teams, for example, are using AI tools to access standard operating procedures across divisions, particularly valuable in an organization that has grown through acquisitions. “There is no need to recreate the wheel,” Cartia noted. “There’s a huge efficiency gain there.”

On the commercial side, AI is helping teams better manage customer interactions. Tools like Salesforce Einstein allow employees to quickly summarize communications and identify next steps.

“Being able to respond to customers faster, that’s something you can tie to revenue,” Cartia said.

At Spee-Dee, even relatively simple applications are delivering value. Navin highlighted a tool that automates service quotation preparation by pulling real-time travel data into proposals.

“It’s more accurate, and it saves time,” he said. “We’re not guessing anymore—we’re giving customers realistic numbers.”

Governance and structure are critical

As adoption accelerates, so does the need for governance.

Saad emphasized that even small organizations need a basic framework for managing AI. “It doesn’t have to be complex,” he said. “But you need to define how it will be used, identify champions, and set some standards.”

That structure is essential for building trust and ensuring responsible use.

Cartia stressed the importance of working closely with IT teams. “You’ve got to make sure your data is protected,” he said. “That’s paramount.”

He also cautioned companies to carefully evaluate third-party vendors. “If a vendor tells you your data is 100% protected, I’d question that,” he said. “And you need to ask whether they’re outsourcing any part of the solution.”

At ProMach, governance includes restricting access to certain tools and creating controlled internal systems. The company has developed its own platform—PAIRIS—to manage AI resources and maintain oversight.

Scaling remains a challenge

Despite growing adoption, scaling AI across an organization remains difficult.

Audience feedback highlighted the primary barriers: lack of expertise, integration with legacy systems, data quality, and unclear ROI.

Cartia noted that expertise often exists within organizations but is fragmented. ProMach is addressing that by identifying internal talent and building a network of AI leaders across business units.

“We’re finding more expertise than we thought,” he said.

Navin said Spee-Dee is taking a more exploratory approach. “We’re encouraging people to use it and come up with use cases,” he said. “It’s about getting comfortable with it.”

Saad added that many companies already have “power users” who can serve as internal champions. The key is identifying those individuals and giving them the tools and structure to succeed.

ROI must be clearly defined

While efficiency gains are widely recognized, the panel emphasized that companies must go further to quantify AI’s financial impact.

Cartia was particularly direct on this point. “You have to tie it to financials,” he said. “It’s very difficult to justify the investment if you can’t show what it means.”

That means translating productivity improvements into measurable outcomes, such as increased throughput or reduced labor requirements.

“If an engineer is more efficient, what does that actually mean?” Cartia asked. “Are we producing more machines? Are we reducing costs? That’s what matters.”

Without that clarity, AI initiatives risk losing momentum as they scale.

While the panelists represented companies at different stages of adoption, they shared a common perspective: AI is not something to wait on.

The most successful approaches are starting small, focusing on real problems, and building from there. Knowledge transfer, customer responsiveness, and internal efficiency are proving to be effective entry points.

For OEMs, the message is clear. AI may still be evolving, but it is already reshaping how companies operate—and those who begin the journey now will be better positioned to capture its long-term value.

As Navin put it, “It’s new, it’s evolving—but you’ve got to get on board to stay relevant.”