June 29, 2026
For independent boxmakers, AI has evolved from future-state concept to present-tense competitive reality
In 2024, AICC’s AI subcommittee released a series of white papers on the promise and challenges of AI for boxmakers. Regarding AI’s potential, committee member and Bay Cities Chairman and CEO Greg Tucker wrote, “AI facilitates a quicker, more accurate rate of knowing what is happening in your plant 24 hours a day. Education and training, office support, and production work is where AI is really going to help.”
At the time, AICC also identified several core challenges affecting that potential, including disconnected systems, data silos, and implementation costs.
Just two years later, AI’s potential is being realized throughout manufacturing. And boxmakers that are using it only expect its role—and benefits—to expand as it becomes a key differentiator within the marketplace.
“Integrating AI tools can allow everything from quicker responses to client emails in sales and customer service to broadening ideation and creative problem-solving in many other parts of our business,” says Jeff Putt, managing director of digital services at DeLine Box & Display and an original member of AICC’s AI subcommittee. “More importantly, when paired with a good business-intelligence (BI) system, it can give us so much more insight into our operation than we ever had before. My immediate goal is to train our people how to use these new tools to make their jobs easier, faster, and more satisfying by removing some of the boring, repetitive labor.”
Still, the challenges identified in 2024 remain. And new ones have come to light, as well, slowing manufacturers’ ability to leverage all AI has to offer.
The Data Foundation
AI’s ability to process information and carry out functions hinges entirely on the data available to it.
“Companies often think AI is plug-and-play, that they can just insert it into a manufacturing process or a document-intelligence solution, and they’re good to go,” says David Wiens, CEO of AI consultancy BPS AI Software. “That’s the first friction they hit: the fact that they need clean data for the AI to do its job.”
“Clean” does not just mean accurate. It means consistent formatting across data types, clear and consistent file naming conventions, and critically, a single accessible data location. In Wiens’ experience, companies typically have “decades of files saved all over the place, in different formats.”
“Typically, an ERP system will control some aspect of their business,” he continues. “Some companies are consolidated into that ERP system with almost everything—accounting, production, sales, all that stuff. Others have ERPs exclusively for financial data and customer service and invoicing but then use an external CRM for maintenance of sales and prospecting pipelines. Then they have SOPs and machine specs and all of these things that are typically saved on some sort of shared-drive location. And then there’s the data coming from machines, which can be a massive amount of information requiring a team of data scientists to pore through, label, and figure out what it actually means.”
As “intelligent” as an AI might be, it cannot work within this scenario. “The solution is to unify your data into a singular source that can be pulled from by machine-learning algorithms and those processes that you need to move the data around,” Wiens explains.
This is no small feat. Depending on the company, just cleaning and compiling data may take three months—or more. Putt says, “We’ve spent the last year working with BI and analytics company Two10 Technologies, who have been consolidating our siloed data stacks into a comprehensive data warehouse. We’ve been able to combine our ERP, floor-shop, maintenance, inventory, quality, and CRM data. That’s typically the hardest part of building an information foundation for AI.”
It may also incur the biggest expense. When AICC assessed the industry in 2024, the entire cost for AI implementation was estimated at $15,000–$100,000. Today, just getting the data in shape for AI implementation—something that wasn’t even a factor two years ago—may cost $150,000–$200,000.
Nevertheless, Putt believes DeLine’s investment is going to be worth the time and expense. “Consolidation will allow us to utilize [data] through a single repository and then use the associated AI to ask it questions about our business processes,” he says. “It is, effectively, a small-language model that uses only our data as its information source. We’ve started to train our ‘content experts’ on how to use this information to make better decisions, reveal trends, and to see if they’re doing better or worse on waste, run speeds, or machine setup times.”
Of course, manufacturers aren’t the only ones turning to AI. Customers are increasingly relying on AI-driven searches to find and vet their suppliers and partners.
Search and Sales
“People have changed the way they buy,” says Todd M. Zielinski, managing director and CEO of Athena SWC. “The generations coming up through the purchasing decision-making level didn’t grow up making deals at dinner or on a golf course. People are busy. They value their time. Because we have AI tools, search, the internet, we can find everything we need at our fingertips.”
“Many salespeople believe the old way is the right way,” Wiens adds. “But I think the responsiveness possible with AI—its ability to respond immediately to inquiries, to have your entire product catalog or custom quoting process available at somebody’s fingertips so they can interact with it—is going to be a differentiator. Once a customer gets used to that ease of doing business, they definitely don’t want to go backward and deal with a different vendor who still needs you to mail a paper PO so somebody can enter it and get you a quote back within a couple of weeks. That’s just not what they’re looking for.”
Because buying habits have changed, boxmakers need to change how they sell. “Most corrugated packagers underestimate the power of their website and inbound marketing to develop lead flow,” Zielinski says, “as well as the speed at which those leads can convert. There’s something missing on the site, whether it’s basic search-engine optimization (SEO) infrastructure, generative-engine optimization (GEO), or just understanding how to write and build content for AI queries.
“There’s a big part of the market that doesn’t understand you can get really good-sized opportunities from leveraging GEO and SEO,” he says. “There’s this outdated assumption that the people at Sara Lee or Nabisco or Walmart don’t use their phones and their laptops to find opportunities for new vendor suppliers and new products and services.” They clearly do. “I’ve read that 75%–80% of industrial buyers—that includes packaging—purchase and make decisions right through search, right through AI tools.”
Zielinski and colleague Lisa Benson, Athena SWC’s senior marketing content consultant, explored the details and benefits of GEO in the March/April issue of BoxScore. But knowing what GEO can do and implementing it for your company are two different things.
Christopher Mays, Athena SWC’s director of operations, says, “When we look at where boxmakers stand digitally now—their websites, their content, their whole online presence—I don’t think most are set up to take advantage of AI. In many cases, their website and what they’re doing online is outdated. That’s why so many of those companies are contacting us for help. We show boxmakers that there are companies searching online for what you do, but your company isn’t showing up. And then we help them to understand the reasons why, so they can make informed changes.”
Among those changes is making sure your website is structurally sound and optimized for SEO in the first place, says Athena SWC Operations Manager Scott Hill. “Not to be overly technical, but that includes making sure you have your address wrapped in an address tag, making sure the schema is set up correctly and is readable for the Google spiders as well as the AI bots.”
To tap into the things GEO is looking for, Wiens recommends posting your product-comparison sheets, case studies with real data, links to articles and blogs that feature your spokespeople. “Things like that aren’t related to keyword searches and whatnot from the old SEO days,” he says. “But they are vital for AI.”
In addition, “FAQ pages are critical,” Benson says. “Include data that’s going to differentiate you from all the other corrugated sites out there. Any factual information you have that is unique is definitely going to help.”
It is also important that the site is featuring answers to questions that customers are actually asking, presented in a clear, straightforward, and factual manner. Those attributes are critical, because AI bots are looking for authority, not marketing copy. “Google’s acronym for this is EEAT: expertise, experience, authority, and trustworthiness,” Hill explains. “The FAQ, customer reviews, trade organization articles where you’re quoted or referenced—all help to establish authority that AI recognizes.”
“There are two different ways potential customers are using AI,” Benson sums up. “If somebody’s doing a Google or Bing search, they will get an AI summary answer with links they can follow. Or they go directly into something like ChatGPT, Claude, or whatever and say, ‘Give me a list of corrugated companies near me and compare them in terms of, say, cost and lead time.’ If that’s the type of data people are looking for, then make sure your website includes it.”
Tomorrow’s Tech at Work Today
If you’re not already integrating AI into your workflow, it may be difficult to understand what, exactly, it can do for your company, where it is best suited to deliver quantifiable results.
According to Hill, AI currently is proving beneficial for manufacturing salespeople: speeding up list development, helping with problem-solving, and drafting prospecting emails. Those that get the most out of it are telling the AI exactly what they need. “The more detailed prompts you can write, the better the results are,” Hill says. “For list development, I might tell the AI, ‘This is who I’m building a list for. Here’s their website. These are the types of industries, the locations, and the size of the companies that I’m looking for as potential customers.’ And I might start by asking for a list of 10 companies that I can review. Once I see the results, I can give feedback: These were good, this one wasn’t because of this or that, and that helps the process. Then we can build on that list to find the complete contact information at the company, and so forth.”
Putt acknowledges that DeLine Box is still in the early stages of using AI tools, but so far AI is having the most impact in the front office and in sales. “We’re using it more to do market, channel, and prospective-customer research,” he says, “as well as to automate repetitive tasks in the office.”
Because AI is not yet fully integrated into DeLine’s operations, Putt is training staff to use enterprise-level AI as a starting point. “It is very important to use the ‘pro’ versions of AI chatbots,” he says, “because they can then be used without the fear of sharing proprietary information.
“Here is a simple example: I’m training salespeople to use agentic AI for trip planning. In agent mode, you can give your destination, travel dates/times, preferred airlines and hotels, and then let the agent mine the best matches and present rates to you. It can also book your flight and hotel.”
While some (especially large) companies may be using AI in recruitment, résumé sorting, and other HR functions, DeLine is not there yet. AI also is not yet helping with predictive maintenance, quality control, production scheduling, and other plant-floor applications. But he believes it’s simply a matter of time. “We are working toward using AI for something much more powerful,” he says.
In the meantime, Putt is incorporating AI into his own workflow. “As an early adopter of ChatGPT, I now use both ChatGPT and Claude daily, and depending on what I need the AI to do, I use several other AI tools, including Midjourney for text-to-image generation, Grok or Runway for AI-created videos.”
Video may not be the first AI application that comes to mind for most boxmakers. But Putt has found it to be very useful. “I recently used the Grok AI to give a bakery a quick idea of the kind of POP display we could develop for them,” he says. “We had developed a four-shelf floor display as a sample, with the customer’s logo and graphics on it. I took a photo of the empty display, uploaded it to Grok, pointed Grok at the company’s website, and said, ‘Fill it with an appropriate product.’ Grok built a 15-second video of someone stacking packaged loaves of bread—with generated SKU labels—onto the shelves.
“The fact that this could be done so quickly means that our sales reps can almost effortlessly go from a concept to a visual mockup and video presentation. If we send a video like this to a buyer or marketing person and say, ‘Hey, we’ve got some ideas for you. Look at this,’ that’s going to be a much more compelling reason to talk to us compared to somebody who emails: ‘We’re selling boxes. You buy any boxes?’ ”
Getting Started
Having a firm data foundation is critical for a comprehensive approach to AI implementation. But that doesn’t mean companies can’t begin benefiting from AI right away. For example, DeLine is launching a contest to help employees get used to using AI through a simple ground-level exercise designed to produce benefits across the company (see sidebar). There’s no reason other boxmakers couldn’t follow the example.
For more widespread implementations, Wiens suggests taking a close look at all of your repetitive processes to figure out where AI could have the most immediate impact. “Ask, ‘Is this something AI does well autonomously? Does it need a human in the loop? Or is this predominantly a human process that AI could just augment?’ Start with the things AI can do really well on its own and the ones with the human in the loop. These will start the process of helping your team learn how to use AI. The activities that AI just augments can come later.”
“Start small,” Putt adds. “Get some little wins, and give your people an incentive to use it. Importantly, start with the enterprise version of your AI of choice: Claude, ChatGPT, or Copilot. Purchase subscriptions for the employees who want to try it, and then give them some kind of incentive to use it.
“The part that’s going to be hardest for many people is moving beyond the ‘I’ll use AI to write a cool email’ response,” Putt continues. “Yeah, it can do that; but where it really matters is getting to the next level, which involves really understanding what your data is telling you about your job and your business. That means you’ve got to get all your data clean and organized—and you probably can’t do that without some help. It’s going to be time-consuming and not inexpensive. But I tell people in our industry, ‘You’ve got to do it. You’ve got to start.’ ”
In fact, Putt points out, there are some corrugated leaders who have been working for years to build agentic systems capable of driving autonomous workflows. “A customer can sign in to a web portal, choose a box design, get a price, order the box, submit artwork, and it’ll go straight to a digital press, get cut on a cutter, and ship out. A week from order, their custom boxes are delivered to their door—all without a human being involved along the way.That means their people can spend their time handling the bigger jobs, the more complicated jobs, and the long-term clients.”
For DeLine—and likely many other independents—this kind of customer-facing, AI-driven workflow automation is still on the horizon. “But it is possible,” Putt stresses, “and coming to a box plant near you. Our business must adapt and learn these new modalities, or we’re going to be left behind.”
Robert Bittner is a Michigan-based freelance journalist and frequent BoxScore contributor.
Sidebar: Artificial Interactions
One of the most difficult aspects of introducing AI is getting people to actually use it.
Jeff Putt, managing director of digital services at DeLine Box & Display, notes, “At DeLine, some of our team members are more excited about working with AI than others. I have found that user resistance is generally a psychological barrier, not a technological one. It’s the intimidation of a new thing and the concern that they might make mistakes: ‘What if I don’t ask the right questions? What if my prompts are stupid?’ What most people don’t realize is that you can have the AI help you write the prompts or even give you advice on where to start with AI. Just type in what you’re thinking or feeling, and it will start helping you figure it out.”
To help people appreciate the possibilities, Putt is launching a contest designed to encourage people to give AI a try. “We will be asking them to use AI to come up with ideas to either increase profitability in our manufacturing operation or to lower cost or waste. We plan to give away a $200 gift card just to get people to engage. Whoever has the best idea wins!”
Putt believes this will be a fun yet productive step to help people get comfortable with AI assistance. “Until people start working with AI, they won’t understand how it can help them in their jobs. It can do a lot more than just help you write an email quickly.”
