AI Adoption Is a Journey. Are You Heading In the Right Direction?

In this industry, transformation is normal — and now a necessity, as Toni McQuilken made clear in this story. It makes sense, then, that artificial intelligence (AI) — which is constantly evolving — would weave its way into the daily work of print service providers (PSPs).

What’s different about this technological innovation compared to a new press or software is that it touches everything and everyone. AI lives in operations providing apparel decoration, packaging converting, wide-format signage, and transactional print. It sits on standby to support employees in production, prepress, customer service, sales, marketing, and beyond.

This pervasiveness is perhaps partly what prompts PSPs to view it as a critical tool. In the Alliance Insights report “AI Adoption in the Printing Industry: From Curiosity to Competitive Advantage,” 85% of print providers surveyed said AI is essential to remain competitive, and 83% said it unlocks new opportunities for them. Tellingly, 42% said AI is now a necessity for survival.

That’s where things were in 2025, when PSPs were largely still testing the waters. Now, some PSPs have begun to move past the experimentation phase, according to Amy Servi-Bonner, VP of PRINTING AI (find the PRINTING AI Pavilion at Booth N6531).

“The center of gravity has clearly moved past proof of concept,” she says. “The ROI is real, but it’s real where it’s targeted. So, the companies that are seeing returns picked one workflow, they’ve measured it, and they’ve expanded from there. And the ones that are still stuck in experimentation are those that tried to boil the ocean or bought a tool before they really understood the problem they were solving for.”

To track how this evolution is showing up in real life, we caught up with Brian Adam and Brian Hite, who shared their AI progress in last year’s PRINTING United Expo Show Guide.

Current State

Brian Hite says Foothill Ranch, CA-based Image Options, where he is principal and co-founder, has expanded its AI use to help with everything from customer presentations and social media to sales and doing rough estimates on big jobs.

But Hite explains that Image Options — a PRINTING United Alliance member — is careful about where it lets AI intervene. With estimation for an architectural project, for example, he says AI can help get in the ballpark of what a job will cost by looking at industry-wide research — about as close as an employee would get.

“Now, the submittal isn’t required to be an actual estimate at that point. It’s just required to be a rough estimate of ‘yeah, this project’s going to cost a million dollars or half a million dollars,’ and that’s what the client is expecting,” Hite said during a panel discussion at the annual Wide-Format Summit (WFS), hosted this July in Bonita Springs, FL. “Now, as you get into the details and start engineering the solutions and doing value engineering on the products, that becomes more of a manual process with experienced estimators.”

Like Image Options, Olympus Group — a PRINTING United Alliance member headquartered in Milwaukee — is using many AI tools across many functions. According to Brian Adam, owner and president, his team has a Claude corporate account, but he also personally employs tools like ChatGPT, Fathom, Replit, Manus, and Limitless.

Leveling Up

Things have progressed past simple tool experimentation for both Adam and Hite, but that doesn’t mean there aren’t new things they are testing out. In particular, Adam has been tinkering with agentic AI.

“We’ve dabbled in agentic AI — creating simplified reports, automating sending them out, and summarizing meeting to-do lists,” he explains. “We don’t have a ton of agents completely built out. We wanted to get proficient at the AI tools that we were involved with so we knew them.”

At Image Options, Hite says he’s been exploring how agents can help expedite the sales research process. In one case, a sales rep of his had been reaching out to a customer for weeks with no progress. So, Hite created a detailed prompt, gave AI certain permissions to research the company, and was able to deliver accurate contact info and other facts about the company to the rep within 15 minutes.

“It cut through all the LinkedIn searches and online Google searches for different contacts from the company to pull information together,” he explains. “It spun up probably 30 different agents based on the query, and each agent came back with its own answer so it wouldn’t corrupt prior answers. … [Our rep] then went out and found the contacts that he really should be working with, because the contacts that he had that he got from LinkedIn had left the company.”

Since then, Hite has turned the function into a webpage where reps can see what companies have already been researched and add new ones to the list. The process runs on Hite’s server and delivers an executive summary, a full report, and a spreadsheet with the contact info it digs up.

Evaluating AI Success

As Servi-Bonner often says, if AI doesn’t hit your P&L, it’s just entertainment. But in order to see financial benefits from an AI initiative, you must measure success. According to Alliance Insights’ “State of the Industry Report 2026,” PSPs who want to get started with AI or push their adoption forward should follow a disciplined four-step framework:

1. Establish baseline performance metrics before implementation.
2. Define a primary objective, which can range from cost reduction to satisfaction improvement, and everything in between.
3. Start with clearly defined, high-frequency inquiries before expanding the scope of automation.
4. Monitor key performance indicators regularly to ensure steady improvement.

On Servi-Bonner’s adoption scale, Olympus Group is beyond experimentation, but still in the process of encouraging adoption across the board. So, Adam explains, ensuring employees are using the AI tools they have is currently a major measure of success.

“Right now, it’s like, OK, how many people are on the Claude corporate account?” he says. “You really don’t want to throw money away. Our objective is we have as many people using it as possible, because once people start utilizing it, it’s going to make them more efficient at what they do.”

However, he notes that success looks different depending on the goal and the AI tool being used to reach it. For instance, Fathom, a tool for recording and summarizing meetings, has been highly successful in Adam’s eyes.

“I know it’s having an impact when I’m involved in more meetings than anybody else, and I know the impact it has for me in terms of my ability to … keep on task and send follow-ups,” he says.

Hite similarly says AI success is currently viewed through a subjective lens at Image Options.

Meanwhile, other PSPs have been taking steps toward the ROI end of Servi-Bonner’s scale. Take Hite’s fellow WFS panelist, Dilo Wijesuriya, president and COO of ARC Document Solutions/Riot Color. At his company, which is a PRINTING United Alliance member with about 100 locations across the U.S., AI is being implemented site by site within its sales team. Now, roughly 100 reps are trained on the process.

“For me, the couple of KPIs I’m looking for is the number of new customers that our sales reps are able to touch, and that is definitely continuing to go up,” Wijesuriya said during the panel discussion. “If they were working with 20 prospects at a time, now that portfolio is a lot larger than before, because now they know about the customers well before they are able to go and meet the customers.”

Even with a decrease in sales reps over the past few years, Wijesuriya notes there’s been nearly an 8% increase in the number of customer touches.

Keeping the AI Train on Track

AI is powerful, but it’s not perfect. The occasional occurrence of hallucinations — i.e., an AI tool fabricates information, often with unearned confidence — means you may sometimes get an answer that’s not quite right. Plus, different tools have different levels of data security baked into them.

But the AI train is not slowing down, so one of the best ways to mitigate issues on this front is to establish an AI governance policy.

Image Options has such a policy in the works. According to Hite, it will outline several guardrails: approved tools and their uses, what data can be inputted into an AI tool, and company-developed tools employees have access to.

Adam says his team began working on theirs around mid-2026. He notes that employees reading and agreeing to the policy “is really a tool — encouraging individuals to use it; encouraging them to play around, try different tools, but do it in a responsible manner; trying to articulate what’s safe and what’s not safe to share, which varies based off of the tool you’re using for us.”

Complementing its governance policy, Olympus Group has organized an AI steering committee, which Adam heads. Where the governance policy sets barriers on responsible use, the committee establishes how AI tools will be chosen, purchased, and reviewed, as well as how the company plans to encourage AI use among employees.

The Next Transformation

Hite and Adam continue to seek ways to strengthen their businesses through the latest AI developments.

Hite has his sights set on creating a solid data foundation. Not only will this support the company’s current implementation of PlanProphet, but it will also make any future AI tools they add more effective.

The plan, Hite says, “is to collect all of the contact information for all of those systems that we work in, collate them all together, and then de-duplicate contacts, correct information — like ‘this person is no longer with the company,’ ‘this person is still with the company,’ or ‘this person is part of the company, but you don’t have them in either of your databases’ — and then verify the emails, verify the phone numbers, the contact information, business addresses, etc.”

Adam has some personal AI projects he’s looking forward to — like vibe coding — but at an organizational level, he wants to expand the use of agents beyond himself.

“I want to get comfortable with agents, so I better understand how to use and deploy agents to better support the team,” he says. “I want to be an advocate for its use within Olympus. I want to put my arms around the employees and show them how it can be a tool that’ll help make everybody more efficient and better at what they do.”