AWS re:Invent 2025 - Modernizing Operational Technology for AI-power Manufacturing (IND370)

Modernizing Operational Technology for AI-Powered Manufacturing

Challenges Facing the Manufacturing Industry

• The manufacturing industry is facing unique challenges that are driving the need for re-industrialization:
  • Continuously building supply chain resiliency
  • Managing the impact of an aging workforce and demographics on business continuity
  • Mitigating risks related to geopolitics and trade

The Role of AI in Re-Industrialization

• Re-industrialization must be AI-driven to build a competitive advantage through manufacturing and supply chain operations.
• The winners in re-industrialization will be those that focus on delivering financial ROI and addressing the durable needs of the operation or supply chain:
  1. Reducing variation and achieving predictable, consistent performance on key metrics
  2. Empowering employees by using AI to enable faster, better decision-making and reduce cognitive load
  3. Executing at speed and reoptimizing capacity to adapt to changing conditions

Challenges with Existing Manufacturing Data and Systems

• Only about a third of the industry has their data in a state that allows for scalable innovation and rapid AI implementation across multiple facilities.
• The traditional ISA 95 Purdue pyramid model is not well-suited to support the requirements of an AI-driven world, with inherent data silos and variation across layers.

A Framework for Modernizing Operational Technology

1. Digitally Composed Outcomes:

   • Redefining common manufacturing and supply chain KPIs (e.g., maintenance reliability, quality, on-time delivery) by integrating AI into the underlying workflows and processes.
   • Moving beyond point solutions (e.g., computer vision, predictive maintenance) to holistically rethink how work is done.

2. Industrial Data Fabric:

   • Driving contextualization of operational technology (OT) data as close to the source as possible.
   • Building a knowledge graph to enable scalable innovation and the ability to replicate successes across facilities.

3. Manufacturing Application Modernization:

   • Transitioning traditional OT applications (SCADA, historians, MES) to a modern, hybrid cloud-based architecture.
   • Maintaining the required latency and security at the edge while leveraging the cost efficiency and DevOps capabilities of the cloud.

Integrating AI into Manufacturing Workflows

• Avoid simply automating existing, broken processes - instead, start with a blank slate to redesign workflows.
• Leverage different AI tools (automation, assistance, chatbots, agents) strategically based on the complexity and risk profile of each task.
• Establish a new baseline for continuous improvement, building on the organization's existing Kaizen culture.

The Role of Leadership in Driving Transformation

• The speed of transformation comes down to a leadership decision to encourage teams to dive in and start improving workflows through AI systems engineering.
• The best place to start is to identify a specific workflow that is ripe for improvement, define the target KPI, and work backwards to determine the data and modernization requirements.