As the field of artificial intelligence (AI) continues to evolve, one thing is clear: Print service providers (PSPs) must start reaping measurable gains from their AI use. That's exactly the message Nathan Safran — VP of Alliance Insights, the research division of PRINTING United Alliance — has for the industry; at PRINTING United Expo 2026, he urged PSPs to look beyond AI experiments with chatbots and content creation — and toward practical gains in productivity.
Safran's presentation drew on Alliance Insights' third annual AI report, which involved a survey of 294 PSPs and 16 interviews. Reflecting on all three reports, Safran said that they've seen a gradual increase in AI use and adoption, so that 38% of respondents in the latest report reporting having deployed AI in at least one business function.
A noteworthy trend Safran highlighted was the number of respondents not experimenting or using AI in any way. In the 2025 iteration of the report, 23% of participants said they had never used AI. In the 2026 version, a stark 0% of respondents reported they were not currently exploring AI.
But while seeing productivity gains should be PSPs main goal with AI, Safran acknowledged that it's going to take time to get there — and that the process is different from any other software PSPs may have used in their careers in the past.
“What's so fascinating about it is, over your career, someone walks into your office to say, 'Here's our new software. You got to learn how to use it. Here's a half-baked user manual. Go figure it out,'” he explained. “It's very intimidating, takes a lot of time, you don't really have anywhere to go to figure it out, and the adoption curve is challenging. What's so fascinating about AI is to learn how to use AI, ask AI.”
Safran expanded on that with a personal example: "I'm building software now with AI, and I do not have a development coding background. And it's been many, many, many, many steps, but every instance along the way [I'm] asking AI, 'How should I be doing this?' and also questioning what it gives me back."
He also challenged the idea that AI is primarily a tool for reducing headcount. In his account of the research, AI’s promise is helping employees move toward more valuable work. For instance, among AI leaders surveyed, 54% reported being able to shift staff into higher-value work.
Instead, he argued that the biggest opportunity is productivity — especially in operational work that can be more complex than drafting emails or marketing copy. One 14-employee label printer used AI to analyze a years' worth of material-spend data and identified $100,000 in material savings. In another instance, a book and catalog printer discovered it had been using multiple methods to track scheduling — a 20-year-old MIS, in spreadsheets, in someone's head. With AI, the company brought this scattered information together in a dashboard, reducing late jobs from 30 to 12.
Those examples, Safran stressed, did not require a large company or a major technology budget.
“I'm not saying there aren't aspects of implementing AI in your organization that would take a budget,” he said, “but to get started and to start to see some of those wins that we're describing here, it does not. The smallest shops in the study got the clearest results with the subscription because of the agility, because of the scrappiness.”
His practical prescription? Start with a specific business problem, make sure the underlying process is sound, and set a measurable target before a pilot begins. Keep people involved in decisions and actions, particularly when AI makes recommendations.
As Safran put it, “Automating a bad step just makes it faster. It doesn't help.”
The message was less about transforming an entire shop overnight than building momentum: Start small, learn, measure the results, and expand what works.
Access the full Alliance Insights report here: "AI in Print: Data Readiness, Technology Alignment, and Unlocking High-Value Use Cases."
And save the date for PRINTING United Expo 2027, taking place Sept. 14-16 in Las Vegas.
Artificial intelligence was used extensively in the writing of this article.
- Categories:
- Artificial Intelligence (AI)
- People:
- Nathan Safran






