Answer extracted from the The Manufacturing IT Podcast — listen to the full episode below.
The traditional hierarchical separation of manufacturing systems—where level two handles controls, level three runs MES, and level four manages IT—is collapsing as IT departments push deeper into the manufacturing execution layer. IT now directly influences MES systems and is extending control down into batch recipes and operational decisions that once belonged exclusively to controls engineers.
This architectural shift reflects a broader change in how manufacturing organizations think about data flow and system integration. Rather than maintaining strict boundaries between functional silos, the industry is converging toward unified data ecosystems where IT, MES, and controls operate as interconnected layers serving a single strategic objective.
What's driving this convergence? Manufacturing leadership is increasingly focused on leveraging data and automation for competitive advantage. As Tom Heckman discusses on The Manufacturing IT Podcast, many manufacturing CEOs are exploring artificial intelligence and advanced analytics to improve product efficiency and profitability—but they cannot deploy these technologies without breaking down the architectural walls that previously isolated IT from the shop floor.
The pressure to implement AI at scale is exposing the inadequacy of the old level-based model. Manufacturing executives want to apply machine learning to production optimization, predictive maintenance, and quality improvement, but legacy system boundaries prevent the clean data flow required for AI success. This tension is forcing a reckoning across manufacturing IT and operations teams.
The challenge isn't simply technical—it's organizational. When IT begins influencing decisions at the MES and controls level, responsibilities shift, expertise requirements change, and teams must learn to collaborate across disciplines that historically operated independently. The industry is still figuring out how to navigate this transition effectively, and implementation approaches vary widely depending on each organization's maturity and technology stack.
Tom Heckman — Founder and CTO at Sepasoft, with over 20 years in manufacturing automation. He began his career in the 1980s at an Allen Bradley distributorship, specializing in data collection and PLCs, later owned an integration business, and founded Sepasoft in 2010 to modernize manufacturing execution. He transitioned from CEO to CTO to focus on the technology vision that drives the company's platform.
One concrete signal of this shift is how manufacturing operations teams are approaching system selection. Heckman notes that companies are now looking for MES and HMI platforms that integrate seamlessly with open data architectures—technologies that reduce friction between IT infrastructure and operational systems. Platforms like Ignition, which Sepasoft builds upon, have gained traction precisely because they allow IT and operations to work with the same technology stack, eliminating translation layers and proprietary lock-in.
The outcome of this reshuffling is still unfolding. While the conversation on the episode explores how Sepasoft's advanced analysis product, launching in the first quarter of 2025, aims to bridge this gap by combining connectors like MQTT and Kafka with analysis engines that both IT and operations teams can understand, the broader industry consensus remains unsettled. What's clear is that the rigid hierarchical model cannot survive the demands of modern manufacturing—and IT's expansion into the MES space is irreversible.
Transparent pricing models that are simple and flat allow companies to easily add new licenses or new users without complex per-tag or per-report licensing, reducing procurement friction and accelerating deployment.
Sepasoft is releasing a product combining connectors like MQTT, API calls, and Kafka with an advanced analysis engine that includes root downtime cause analysis and other AI-ready features designed for manufacturers.
Many manufacturers struggle with getting clean MES data into AI systems. MES data isn't static after initial recording—quality results change, tests run, and historical data evolves, requiring continuous curation and validation.