Interfilter Group has built its reputation on a straightforward idea: customers should never have to worry about what’s behind their products.

The Netherlands-based wholesaler of air, liquid, gas, and dust filtration solutions has spent more than 45 years making sure the air stays clean, the liquids stay pure, and the industrial environments its customers depend on keep running. Reliability, flexibility, and fast supply are operational commitments, renewed with every order shipped.

As the company approaches its 50th anniversary, those commitments are intact. But honoring them required an honest reckoning: the internal processes behind the promise were starting to crack under the weight of their own success.

When order volumes grew beyond what any team could absorb manually, Interfilter chose technology to elevate the team’s productivity. Partnering with Mendix implementation partner, Jump, they built an AI-powered order portal designed to handle growing demand without disrupting how customers already worked.

Phase 1 went live in four months. Today, 30% of all orders flow through the portal automatically — with a Phase 2 target of 80–90%. Alongside the AI Order Portal, Interfilter also runs the Sticker App, a warehouse labeling solution that marked the company’s first step into the Mendix ecosystem.

Together, they signal a shift in how Interfilter intends to compete by building smarter systems that let every customer feel the difference.

When Growth Outpaces Your Processes

The inbox told the story. Every day, orders arrived in whatever format customers happened to use. No consistent structure. No standard fields. Just a growing stream of requests that someone had to open, interpret, and key in by hand, with no visibility into who was working on what and no way to prevent the same order being processed twice.

“The volume of orders is quite big,” said Jesse Fok, Operational Manager at Interfilter. “With human hands, we couldn’t handle it anymore. That’s the reason we decided to pivot and leverage AI.”

The solution had one non-negotiable constraint: customers could not be asked to change a thing.

“Their goal was to make the ordering as easy as it already was for their customers,” explained Mike Stijnis, a project manager and low code team lead at Jump. “We identified the most time-consuming steps and focused on automating those with AI.”

The platform decision was less a debate than a natural conclusion. Interfilter had already built its first Mendix application — the Sticker App — through a third party. When Jump took over its management, a working relationship quietly took root. When the AI Order Portal need emerged, Interfilter already knew what the platform could do, and they already trusted the team holding the keys to it.

Jump’s eight-person team of Mendix-certified specialists, led by a certified Mendix Expert (and MVP), brought pre-built templates that could accelerate the foundation. And crucially, Mendix’s flexibility meant that changing requirements mid-build would not derail the timeline. On another platform, that might have meant delays and difficult conversations. On Mendix, it meant an afternoon. For a company that had never built anything this ambitious before, that mattered enormously.

“With Mendix, you can easily change requirements without a lot of extra hours,” Stijnis emphasized. “That was incredibly valuable here.”

Building the AI Order Portal

Interfilter’s customers had their own ways of placing orders, and changing that was never on the table. Some sent structured PDFs. Others sent Word documents, Excel files, or plain-text emails. The formats were as varied as the customer base itself, and the portal had to absorb all of them without asking anyone to adapt.

What that meant in practice was a centralized inbox where ten users could log in simultaneously, pull from a shared queue, and work side by side without the chaos of the old process. To prevent two people from inadvertently processing the same request, Jump built an order-locking mechanism: the moment one team member opens an order, it becomes unavailable to everyone else. A small feature, but one that solved a real and recurring problem.

The AI layer is where the heavy lifting happens. The portal is built around two complementary technologies:

1. OCR to extract and read text from incoming documents regardless of format.
2. An OpenAI model to interpret the meaning of what was extracted, identifying product specifications, quantities, and customer details.

The combination was not the original plan. “Early development relied on OpenAI alone, but the sheer variety of document structures quickly exposed its limits,” Stijnis shared. “OCR is much better at understanding text, so combining it with OpenAI gave us the accuracy we needed.”

To push accuracy further, Jump built customer-specific templates for Interfilter’s ten largest accounts. The result: 30% of all orders now flow through the portal automatically, with ten users processing approximately 50–60 orders per day.

“The number of orders that have gone through the app — that’s our biggest moment so far,” Fok said.

Meeting Demand with Speed

The Sticker App tells an equally important story. When an ERP migration rendered Interfilter’s existing warehouse labeling tool unusable, a replacement had to be built fast. The Mendix application that replaced it generates accurate shipping labels, covering customer-specific details, project references, and private label requirements at the point of fulfillment.

“With the Sticker App, that’s one of our biggest service level improvements at the warehouse,” Fok added. “We can supply customers more quickly.”

Built Together, Improved Together

The AI Order Portal did not arrive fully formed. It was shaped continuously by the people using it.

Interfilter’s operations team shared feedback throughout the build, not at formal checkpoints, but as real-world use surfaced real-world friction. Jump responded quickly, adjusting requirements without significant rework. The order-locking mechanism itself came directly from that feedback loop.

Marjolein Hermans, Finance, HR & IT Manager at Interfilter, saw the value in what the platform could accommodate. “It’s a good app because it’s custom-made,” she said. “At Interfilter we have a lot of exceptions, and that makes the app a good solution for us.”

“There was a lot of feedback between the users at Interfilter and Mike’s team,” Fok added. “We shared everything we saw working with the app, and they made solutions. It still works great.”

Value Realized, and What Comes Next

Phase 2 of the AI Order Portal is now underway, and the ambitions are significantly larger:

• All document types supported: Word, Excel, plain text, images
• Unlimited customer prompts, giving Interfilter’s team the ability to configure templates themselves
• Quotations added alongside orders, expanding the portal’s scope beyond order processing
• Automation target: increasing from 30% to 80–90% of all incoming orders

“The ability to create prompts independently for any customer will be a game changer for us. That level of autonomy will give us even more wins,” Fok stressed.

Beyond the portal, transportation planning and warehouse stock management sit on the longer horizon. For now, the focus stays sharp.

“You have to introduce AI slowly,” Stijnis said. “Now that we have a solid foundation, there are more opportunities to embed intelligence into their operations.”

The question Interfilter started with was operational. The answer pointed somewhere bigger.

“We can supply our customers more quickly than before. The main focus was using AI to elevate our team’s productivity in an intentional and strategic way. Mendix made it possible for us to find the right places to embed intelligence without sacrificing our customer experience,” Fok concluded.