Over the Air · 22 Jun 2023 · 26m
Erik Udstuen, Co-Founder & CEO at TwinThread, explains how digital twins optimize manufacturing by integrating sensor data with AI models to improve quality and energy efficiency. He outlines three critical conditions for success: having sufficient sensor data, clear economic value, and collaboration between OT and IT teams. Udstuen shares a case study of a pet food manufacturer using digital twins to automatically adjust ingredients for consistent quality, and discusses the shift toward autonomous operations and global 'virtual centers of excellence' driven by labor costs and supply chain resilience.
Where this conversation describes AI being used to build or deliver a product. Part of the AI in practice tracker.
ManufacturingTwinThread platform
AI models analyzing sensor data from a pet food production line automatically manipulated ingredients in the control system to maintain optimal quality despite natural variation in raw materials.
“models for controlling quality, and those models produce recommendations that get sent all the way back down to the control system to automatically manipulate ingredients on that production line to get the most optimum quality out the end.”
Speaker 1 · Hear it at 3:06
companies scaling up
Other shows on the same ground, closest first.
Daniel First, founder and CEO of Axion, explains how AI is transforming quality and innovation roles in manufacturing by turning overwhelming streams of customer, device, and service data into actionable insights. He argues that quality teams are shifting from plant-floor oversight to strategic partners across engineering, marketing, and supply chain, using AI to accelerate product evolution and break down silos. First highlights specific examples, such as Shark Ninja’s ice cream machine and power systems in mines, to illustrate how unexpected customer usage drives innovation and how AI agents can detect root causes faster than humans.
Ian Burns, President of Applied Control Engineering (ACE), discusses the transition of midsize manufacturers from paper and spreadsheets to connected, data-driven environments. He explains how digital twins are used for operator training and control system validation to reduce downtime, and highlights the critical need for Operational Technology (OT) cybersecurity as systems become interconnected. Burns also notes that despite economic uncertainty, manufacturers can find value in incremental data investments rather than just large capital projects.
Anna-Katrina Shedletsky, CEO and founder of Instrumental, discusses how the pandemic accelerated digital transformation in electronics manufacturing, validating her prediction of five years of innovation in 18 months. She explains Instrumental’s cloud-based platform that collects engineering data from across the supply chain to identify defects and improve first-pass yield, moving away from opaque, just-in-time models toward resilient, data-driven collaboration. Shedletsky highlights the shift from Capex to Opex for manufacturing technology and the critical need for transparency between brands and Contract Manufacturers (CMs) to solve problems faster.
Sam George, Board Member of Cosmo Tech and former Corporate VP of Azure IoT at Microsoft, explains how digital twins and AI-guided simulation help enterprises meet KPIs and demonstrate ROI. He contrasts simple operational awareness with complex goal-seeking simulations, using Michelin’s supply chain optimization as a primary example of reducing carbon emissions and increasing profitability. George emphasizes that deterministic generative AI guides these simulations to find optimal outcomes in complex environments, democratizing access to tools that were previously the domain of advanced experts.
Scott Everett, Co-Founder and CEO of Eigen Innovations, explains how his company uses industrial vision systems and AI to transform manufacturing quality control from reactive, destructive testing to proactive, insight-driven decision-making. He argues that while IoT infrastructure for data capture is mature, the industry struggles with 'insight architecture,' requiring contextualized data to provide explainable stories that empower non-technical operators to make real-time adjustments. Everett highlights specific applications in the automotive sector, such as ensuring the hermetic seal of vehicle headlights, where 100% inline inspection replaces random sampling to achieve zero-defect manufacturing.
The summary, chapters and topics are generated from the episode’s audio by this directory, not written by Over the Air. The episode itself belongs to its publisher.