IPMA Report: Be Proactive About AI

The 2026 In-plant Printing and Mailing Association conference brought nearly 130 managers to Greenville, South Carolina. After four days of educational sessions, networking, and awards, attendees returned home excited to implement some new strategies and tools to improve their operations.

One intriguing general session was led by Amy Servi-Bonner, vice president of the PRINTING AI division of PRINTING United Alliance. Throughout her career, Servi-Bonner has guided companies through periods of immense technological change, such as the introductions of e-commerce and ERP. Now, she is helping the printing industry navigate the quickly-shifting artificial intelligence (AI) scene.

After witnessing all this change, her message to attendees was clear: In-plants cannot afford to wait and see how AI turns out.

“[Employees] are not waiting for some policy to come down,” she said. “They're already figuring out how to use free tools to get the job done, because that's what good employees do. … Every week that you all wait to start getting into AI, the vacuum gets filled, and the people that are filling that right now are the people in your shops — without governance, without protection, and without anyone asking whether that's the right approach.”

To combat this, Servi-Bonner presented the framework of rewire, build, and scale.

As part of the rewire phase, she urged in-plants to be proactive in their organizations when it comes to AI. Since in-plants hold and manage data, they have every right to start the conversation on AI governance within their organizations.

In the build stage, shops must identify at least one workflow in which a governed AI assistant would make a difference, such as brand compliance, mail prep exceptions, and job estimation. Once an area is identified, in-plants can start the process of creating an AI assistant that contains the institutional knowledge trapped in their most critical employees’ heads. This can help shops retain decades of expertise as some of their longest-tenured workers start thinking about leaving.

“Here's the thing that most operations don't think about until it's too late: This knowledge doesn't leave when the person retires,” Servi-Bonner said. “It starts leaving the moment that they've decided that they're done, the moment they stop caring about solving problems at seven in the morning. And that's why the build part matters.”

Finally, it’s time to scale. This is not necessarily about building more, but about producing small, measurable results over time.

“If you bring a small, governed AI proposal to IT, IT’s going to say ‘yes,’ because it's specific, it's contained, it maps to data classification that they already manage,” Servi-Bonner said. “The pilot runs, it produces a result — a measurable one, even a small one — and then six months later, when the enterprise AI committee is figuring out which departments to include in phase two, the in-plant can already be in the room. Not because anyone advocated for them, but because they had a result.”

The IPMA 2026 conference was filled with insightful sessions like this. We’ll bring you more highlights in the days ahead.