FESPA 2026 Recap: AI in the Corrugated Cardboard Industry
Written by Thomas Othax, CEO, Packitoo
Last May, I had the privilege of speaking at the first edition of the Corrugated Conference at FESPA 2026. I had been asked to address a topic that seemed simple on paper: AI as a solution to support the automation and digitization of the corrugated cardboard industry. But as I prepared my presentation, I realized that the real question wasn’t whether AI is the solution—but why, in our industry, we still need to be convinced that automation and digitalization matter. So that’s where I began.
Your business model needs to evolve
If you run a corrugated box manufacturing plant or a cardboard packaging business today, the pressures are very real. Squeezed margins, rising raw material costs, estimators, sales representatives, and—more broadly—senior experts retiring and taking thirty years of expertise with them. At the same time, you probably have two sales representatives who quote the same project in two different ways, with bottlenecks at every approval stage. You have to manage manual data re-entry between systems that should have been communicating with each other for the past ten years.
In a world where everything happening outside our walls is accelerating and everything happening inside takes time, the honest answer to the question “Can we wait?” is no. This perspective is essential: if AI is merely a trendy layer applied to flawed processes, it will amplify the dysfunction rather than resolve it. The starting point isn’t AI—it’s the realization that your operational model needs to evolve.
From Rules to Reasoning, Then to Action
When I explained agentic AI on stage, I used an analogy from *The Matrix* that really resonated with the audience. Remember the scene where Neo steps out of the Matrix, sits down in a training chair, and someone loads the “kung fu” skill onto him? The main character—in his training matrix—then instantly declares, “I know kung fu.”
To put it simply, that’s exactly what a *skill* does for an AI agent. Give it the skill “structural designer specializing in corrugated packaging,” and once it’s loaded, the agent now understands cardboard properties—such as ECT, RCV, the COBB index, packaging and palletizing constraints, and so on… It speaks the same language, understands the physics behind every manufacturing decision, and applies this logic to its own set of rules. And when you load a different *skill*—for example, “pricing expert”—the Agent becomes someone else, enriching the knowledge base you initially defined in the process.
That’s what agent-based AI delivers: a combination of automation (workflows and tools), LLM (reasoning), and *skills* (domain knowledge). Together, these elements create an agent that understands your business, leverages your data, and can even act on your behalf.
Deploying AI in the corrugated cardboard industry: a challenge in its own right
This is a classic case of confusing correlation with causation, and this, I believe, is where most industry discussions go wrong. AI in the corrugated cardboard industry is not the same as AI in retail or finance. The constraints are just as demanding, the variables are different, and the cost of error is higher.
In an industry facing an increasingly restrictive regulatory environment, “almost right” isn’t good enough. A general-purpose model will be able to easily generate an answer regarding the contents of a carton. But an Agent that slightly miscalculates the RCV for palletizing a package will specify the wrong product, resulting in a non-compliance issue and potentially causing you to lose a customer order. In the corrugated cardboard industry, “almost right” is worse than not producing a response at all.
General-purpose AI is a cross-functional tool in a world of specialized problems
Let’s take a concrete example. ChatGPT doesn’t know that a 175-gram unbleached kraft liner on B-flute behaves differently at 85% humidity versus 40%. It doesn’t understand that your customer accepts a tolerance of ±2 mm on the interior dimensions, but your competitor does not. A vertical AI designed for the packaging industry doesn’t just speak your language—it also understands the “how” and “why” behind packaging constructed in this way, and it incorporates rules specific to each client.
As I said: AI doesn’t fix flawed processes—it scales them. So, if your quote goes through ten people, across four different systems, and takes two business days to put together, adding an agent to that same scenario will only result in a faster version of the same problem. The companies that derive real value from AI are those that have first rethought their workflows.
Knowledge lives within you, within your team, and within our industry
Where the Real Opportunity Begins
A recent article on WhatTheyThink by Pat McGrew and Ryan McAbee puts it well. In the printing and packaging industry, the most tangible AI gains don’t happen on the production floor—they happen much earlier in the process. That’s where sales teams, customer service, estimators, and administration turn the customer’s intent into a production reality.
That’s exactly what we’re building at Packitoo: AI agents, each with its own role and function. The first is a business analyst who supports the sales team by identifying untapped opportunities and preparing for client meetings. The second is designed as a technical expert—the digital twin of your senior structural designer. The third is a pricing strategist who highlights the profit margins you’re leaving on the table with every quote. The sales assistant is the fourth agent: they turn an email or attachment into a draft quote. And finally, the project manager is the fifth agent—they ensure that deadlines and approvals stay on track.
What You Can Do as a Leader in the Corrugated Cardboard Industry
I concluded my presentation with a simple framework that you can start building right now. The well-known principle of “writing what you know” applies fully here. You can learn by doing, audit your processes, and align AI with your strategy. In the first twelve months, you can connect agents to your ERP and MIS data in real time, deploy your first use cases with pilot users, and build feedback loops. Beyond that first year, you can work on deploying agent-based capabilities across your operations and embedding institutional knowledge in your systems—not just in your employees’ minds.
The greatest risk isn't failure. The greatest risk is not taking action at all.
You can also watch this video to see my presentation at FESPA.
Interested in learning more about our new agents?