May 7
Loading video player...
Designing Multi-Agent Systems for Industrial Operations: Kence Anderson, CEO & Founder of AMESA, joins Kudzai Manditereza on the AI in Manufacturing podcast to explore how machine teaching methodology and multi-agent systems are transforming how manufacturers capture expert knowledge and deploy autonomous decision-making on the plant floor. Podcast links: • Spotify: https://open.spotify.com/episode/2x9eK31xvxuDfxbDvuqrNO?si=27b292c224554823 • Apple: https://podcasts.apple.com/de/podcast/industry40-tv/id1541197224?l=en-GB&i=1000766579539 ## Timestamps 00:00 – Introduction 02:55 – The Biggest Gap Between AI Promise and Reality in Manufacturing 05:39 – The Research-to-PR Pipeline: Why R&D's "Development" Phase Is Missing 08:30 – Real-World Example: Teaching AI to Run a Cheetos Extruder 10:47 – Pilot Purgatory and Misallocated Innovation Bandwidth 12:21 – Four Ways to Make Decisions: Calculate, Search, Look Up, or Learn by Practicing 15:28 – What Is Machine Teaching and How Does It Differ from Traditional AI Approaches 19:30 – Glass Manufacturer Case Study 24:42 – Why Multi-Agent Design Beats Monolithic AI in Manufacturing 29:05 – Multi-Agent Design Patterns: Strategy, Perception, Plan & Execute 34:49 – Digital Twins vs. Agents: Separating Feedback from Action 37:05 – AMESA Platform Overview: Agent Orchestration Studio, Agent Cloud, and Runtime 40:02 – Building Multi-Agent Systems Without Code in the Agent Orchestration Studio 42:50 – Simulation-Driven Training: Data Scores, Simulation Scores, and Faithful Digital Environments 45:30 – Deploying Agents to Production: Edge Runtime, PLCs, and Industrial Protocols 47:53 – Scaling AI Agents Across Plants, Recipes, and Formulas Using Operating Regions 50:49 – Real-World Customers: Chemical Process Control, Logistics, and US Military 52:42 – From Data to Autonomy in 12 Weeks: How to Engage with Amesa 54:03 – Prediction: The Bifurcation of Gigafactories vs. Human-Agent Collaboration AI in Manufacturing Podcast with Kudzai Manditereza ***************** Connect With Me: • LinkedIn: https://www.linkedin.com/in/kudzaimanditereza/ • Website: https://www.industry40.tv/ ***************** Guest - Kence Anderson ***************** • LinkedIn: https://www.linkedin.com/in/kence/ • Website: https://www.amesa.com/ Kence Anderson is the founder and CEO of AMESA and former Director of Autonomous AI Adoption at Microsoft. He is a pioneer in the field of intelligent autonomous agents, having co-created “Machine Teaching”, a methodology that enables AI agents to develop real-world autonomy through simulation, feedback, and trial-and-error. Over the past seven years, Kence has focused exclusively on designing, building, and deploying intelligent autonomous agents for manufacturing and logistics, leading over 200 real-world deployments for major corporations, including Shell, PepsiCo, and Delta Airlines. He is also the author of Designing Autonomous AI (O’Reilly, 2022) and is now developing a horizontal platform for orchestrating AI agents to make million-dollar decisions in enterprise operations. ***************** PODCAST SPONSORED BY ***************** HiveMQ: https://www.hivemq.com/ *****************