In 2021, Vivix Vidros Planos, one of Brazil’s leading flat glass manufacturers, started a bold journey with Mendix. The goal was clear: streamline complex operations, accelerate innovation, and equip frontline teams with smarter tools, from full product traceability to real-time process control.

“We rely on digital solutions to bring multiple forms of value to our company,” shared Aristóteles Terceiro Neto, Industrial Transformation Manager at Vivix. “While ROI is a key goal, we also aim to improve the daily work experience for our operators and managers.”

Since 2024, Vivix’s pioneering approach has earned global recognition, including the Siemens Texcellence Award and the AWS Gen AI Gamechanger.

At Realize Live Detroit 2025, Neto shared how the team rapidly scaled its Mendix portfolio, now almost 30 applications strong, while building a resilient, AI-ready data ecosystem.

“One of the industry’s continued challenges is normalizing and using data effectively,” Neto said. “We needed to integrate this data with Mendix to rapidly scale our solutions and develop different apps tailored to our needs. This requires us to carefully balance effort, cost, and performance.”

Today, Vivix is moving beyond digital transformation and into intelligent operations where advanced analytics and AI drive continuous improvement in manufacturing performance.

Advancing Industrial Digital Maturity

At Vivix, it became increasingly evident how complex and fragmented industrial data can be. The company dealt with a wide range of information like engineering diagrams, 3D models, continuous machine alerts, and extensive production records.

Most of this data was originally designed to control equipment, not to be analyzed. Harmonizing and integrating these diverse sources into a cohesive, usable system was one of their greatest challenges.

“Most industrial companies take a lot of time to get in data processing daily,” Neto emphasized. “Normally, it’s about 80% of the time in the force do it.”

Inconsistent data formats, units, filters, and collection intervals – often across identical equipment – along with unsynchronized timestamps between PLCs and backend systems, hindered scalable, reliable manufacturing solutions.

Building on their new Mendix foundation, Vivix decided to implement a DataOps architecture designed to connect Mendix with real-time shopfloor data from their disparate OT systems, including Industrial Edge deployed application data.

They introduced HighByte on AWS as a central data modeling and integration layer, enabling them to organize industrial data from three key perspectives: assets, processes, and products.

Establishing Core Data Structures

Embracing the industrial data fabric approach, Vivix prioritized the comprehensive storage, contextualization, and utilization of data throughout the entire organization.

By integrating Industrial Edge with cloud infrastructure then bridging the shop floor to enterprise systems, Mendix served as the unifying application layer that enabled real-time insights and informed decision-making.

In 2025, the company further expanded its digital ecosystem by integrating partners like Snowflake and Neo4j to further enhance data capabilities.

Yet, challenges remained. Fragmented and siloed data often produced incomplete or misleading outputs from AI copilots, limiting their effectiveness on the shop floor. Traditional BI dashboards, with static views, struggled to deliver deeper, actionable insights.

Generative AI changed the game by providing instant access to context-rich information and even guiding teams to ask the right questions.

Today, their data architecture consists of:

• Snowflake for scalable, structured analytics
• Senseye, connected to HighByte, bridging shop floor and cloud
• Multiple formats (P&IDs, 3D models, engineering docs) incorporated into layers
• Altair AI to create intelligent solutions seamlessly with Mendix

This marked a significant shift from Industrial Data Fabric to Industrial Knowledge Fabric where data is truly understood and activated.

Throughout their journey to enterprise DataOps and AI digitalization, Vivix kept Mendix front of mind.

With a strategic approach to smart Mendix development, the team began building core applications leveraging artificial intelligence to enhance machine maintenance and operational excellence.

Enhancing Furnace Monitoring with AI

As part of its ongoing digital evolution, Vivix is undergoing a fundamental shift in how it applies AI, moving from reactive systems to proactive, contextual intelligence.

Over its four-year journey with Mendix, the manufacturer has built a growing portfolio of industrial applications including Digital Shift Manager, Glass DNA, Quality Management, Advanced Planning and Scheduling (APS), and Virtual Engineer AI assistant.

But its flagship initiative, Smart Furnace Monitoring, represents a new era of AI-enabled manufacturing.

“Imagine a $120 million furnace, and your job is to make sure its inner walls don’t collapse over the next 18 years,” Neto desribed.

Smart Furnace Monitoring combines real-time industrial data, predictive modeling, and historical analytics to predict furnace degradation before it happens. It bridges Mendix applications with HighByte’s data layer, enabling contextual intelligence across the plant.

To optimize furnace maintenance and extend asset life, Vivix adopted a two-step approach:

1. Simulations and thermal data processing powered by advanced algorithms assess degradation patterns and project future wear.
2. A generative AI-powered Mendix management app delivering actionable insights to maintenance teams and supporting diagnostics, prognostics, predictive maintenance, and knowledge management.

Thermal imaging alone, however, proved unreliable for detecting degradation. In response, Vivix developed a Mendix-based mobile app for field technicians using Fluke devices and smartphones. This AI-powered solution applies computer vision to:

• Pinpoint the exact location of each thermal image
• Auto-correct positioning discrepancies between operators
• Analyze airflow velocity, a key predictor of degradation
• Remove visual obstructions for clearer furnace wall analysis

This combination empowers Vivix to manage one of its most critical assets with greater safety, accuracy, and confidence. For Neto, keeping digital eyes on industrial equipment keeps them operating 24/7.

Most notably, this entire generative AI system was built by Mendix specialists.

Thanks to their streamlined, low-code approach, Vivix is already planning to extend AI further into the Mendix ecosystem, automating workflows like material procurement and maintenance scheduling to:

• Increase asset lifetime
• Enhance operational safety
• Drive data-driven decisions
• Improve asset knowledge management

Powering Digital Transformation at Scale

In just five years, Vivix has pushed the boundaries of industrial innovation and low-code development with Mendix powering every step of the transformation. Evolving from raw data to actionable knowledge and from disconnected systems to an integrated ecosystem, Vivix is running smarter, faster, and with a lot less friction across every part of the operation.

And the journey is far from over.

“Our next goal is transitioning from a digital shadow to a true digital twin,” Neto said. “Now, with Mendix in multi-agent strategy, we can run applications seamlessly across Industrial Edge and cloud.”

Yet, as digital capabilities accelerate, Neto also stressed the critical importance of keeping people at the center of AI transformation. To help spread the benefits of digitalization across Brazil’s industrial sector, Vivix even launched Maya, a regional industrial alliance focused on sharing technical knowledge and digital best practices.

“Our goal is to extend digital transformation lessons learned at Vivix to nearby companies, creating a stronger ecosystem,” Neto said.

Vivix also understands that successful innovation isn’t just technical, but collaborative. Recognizing the investment required for meaningful transformation, Neto emphasized the need for partnerships and scale.

“Digital transformation is not cheap. To go state-of-the-art with AI, you need economic power, and you need knowledge from different companies to create new solutions together,” Neto said.