223: The State of Automation Halfway Through the Decade, Live from the 2025 A3 Business Forum

Manufacturing Happy Hour · 11 Feb 2025 · 1h 12m

AI in practice: 4 real uses of AI described

What this episode covers

Chris Luecke and Jake Hall interview five industry leaders at the 2025 A3 Business Forum, including Peter Zafiro of LinMot USA, Christelle Keefer of ABB, Juan Aparicio of Reshape Automation, David Dechow of Motion Automation Intelligence, and Alex Shikany of A3. The guests discuss the convergence of robotics, AI, and motion control, highlighting how LinMot’s linear motors enable washdown applications in food processing and how Reshape Automation uses deep learning for bin picking. They also address the critical need for early STEM education through organizations like First Robotics and Project Lead The Way to bridge the skills gap, while predicting that physical AI and explainable machine vision will define the next five years of automation.

AI in practice

Where this conversation describes AI being used to build or deliver a product. Part of the AI in practice tracker.

Engineering & simulationCaddyreducing drawing search time by 90% and procurement costs by up to 10%

The tool analyzes engineering drawings to extract data and links them to purchase orders or spec sheets, allowing teams to identify common parts across catalogs to optimize production and reduce costs.

“Caddy analyzes drawings to extract data and creates a link to every PO or cam file or spec sheet. That design has ever been associated with. This lets your team start to leverage hidden commonalities across your parts catalog to optimize production and reduce costs.”

Speaker 2 · Hear it at 30:35

Engineering & simulationReshape Automation / Deep Learning

The company applied deep learning to bin picking by training a dataset of potential grasps for infinite objects, allowing robots to learn how to grasp items without traditional complex programming.

“So if we, instead of that, train a data set of potential grasp for infinite objects, and then learn, then can we add some, um, prolistic methods on the learning into the mix, we work better. Bean picking was perfect.”

Speaker 8 · Hear it at 34:29

Quality & testDeep Learning / Machine Vision

Deep learning algorithms are used for quality inspection by analyzing pixels to detect defects like scratches or cracks, replacing traditional deterministic vision methods that lacked explainability.

“Been picking being run. I would say, my team vision applied to quality inspection. Okay, would be another one. It's still very deep learning bass. So, very learning a steal on pixels. There is no explainability, right?”

Speaker 8 · Hear it at 39:11

Customer supportLLMs (Large Language Models)

LLMs are used as AI co-pilots for human-in-the-loop support, where the AI provides answers to queries that are then curated by humans, enhancing capabilities without the risk of hallucinations in direct robot control.

“But if there is a human giving a a query, then the AI giving you an answer and then another human curating that, the possibilities are endless, right? So very optimistic of doing that as an AI copilot, right? As a human compiler.”

Speaker 8 · Hear it at 39:11

Mentioned

Companies
LinMot USA, ABB, Reshape Automation, Motion Automation Intelligence, A3, Skillwork
Also
First Robotics, Project Lead the Way, REC Foundation

Other shows on the same ground, closest first.

  • The Robot Report Podcast · 1h 28m

    Automate 2025 Recap

    Hosts Mike Oitzman and Gene Demaitre recap the 2025 Automate show, highlighting the shift toward mobile manipulation, AI integration, and ceiling-mounted mobility. They feature interviews with Ujjwal Kumar of Teradyne Robotics on the UR 15 launch, Marc Segura of ABB Robotics on the spinout and AI strategy, Carol Franklin of A3 on new humanoid safety standards, and Mathias Entenmann of CeiliX on their novel ceiling mobility device.

  • EMS@C-LEVEL · 5m

    Electronica 24: The Rise of AI-Powered Robotics at Fideltronik and Beyond

    Piotr Owczarek, President of Fideltronik, discusses the company's seven-year journey to develop AI-driven robotics for electronic manufacturing services (EMS). He explains how a team of 170 engineers in Krakow is creating flexible automation that mimics human operators to address labor shortages and cost pressures. Owczarek details their 'academy for robots' concept, where collective knowledge is stored in the cloud to allow robots to learn from each other without daily programming. The episode highlights that major EMS companies are currently engaged in pilot programs to validate this technology before full-scale adoption.

  • Over the Air · 39m

    Embedded Systems and IoT Connectivity: Lessons from Automate 2024

    Ryan Prosser, CEO of Very Technology, reports from Automate 2024, interviewing leaders from Machine Metrics, Unit X, Sensory Robotics, On Logic, Light Guide, Elementary, Snowbottics, Acme Manufacturing, Happily, Kongitech, MyTech, Lincode, Wheel Me, and Orbit 3D. The conversation covers AI-driven vision platforms, robotic safety sensors, rugged edge computing, and autonomous mobile robots, highlighting the shift toward edge AI and human-centric automation. Key technical discussions include point cloud generation for safety, modular industrial PCs, and AI-based defect detection for high-speed manufacturing lines.

  • The Robot Report Podcast · 1h 05m

    Recap: Promat and GTC 2025 show highlights

    Host Steve Crowe and editors Eugene and Mike recap their experiences at ProMat and NVIDIA GTC 2025, highlighting the shift toward handling chaotic real-world environments in logistics automation. They discuss specific product demonstrations from Agility Robotics, Locust Robotics, and Boston Dynamics, while also covering NVIDIA's announcements on AI simulation and humanoid robotics at GTC.

  • The Robot Report Podcast · 1h 35m

    The Road Ahead for Robotics: Insights from Motional's Major and Foundation's Pathak

    Laura Major, CEO of Motional, discusses the company's autonomous vehicle strategy, emphasizing the shift from technical to business leadership, the use of AI for long-tail problem solving, and their safety record of zero at-fault accidents over 2 million miles. Sankaet Pathek, founder and CEO of Foundation, explains his company's rapid 13-month journey from idea to deployment of a humanoid robot for manufacturing, highlighting their proprietary actuator design and deep variational Bayesian filters (DVBF) for AI. The episode covers the strategic choice to target structured industrial environments to accelerate data collection and the challenges of scaling production to achieve sub-$20,000 unit costs.

The summary, chapters and topics are generated from the episode’s audio by this directory, not written by Manufacturing Happy Hour. The episode itself belongs to its publisher.