Will AI Agents take Manufacturing Jobs or Add Value? | Bryan Debois – RoviSys

January 10, 2026

AI Agents - Bryan Debois

Bryan DeBois is the Director of Industrial AI at RoviSys. He is a fountain of knowledge surrounding the topic of AI. They are working to enable digital transformations for manufacturers throughout the country through the implementation of AI agents.

This conversation revolves around the differences between generative and autonomous AI and the best applications for each. He talks about the inner workings of these different AI models, how his company is using it to fight back against manufacturing brain drain, and the strategies that can help businesses begin to implement the infrastructure needed to fully take advantage of this technology. The most interesting, being their process of interviewing retiring industry experts, training agentic AIs on that interview and machine, and using these AI agents to create workflows to guide and train the next generations of manufacturers.

Nate Wheeler is the owner of weCreate, a nationally recognized marketing agency that helps manufacturers grow, save money, and become more efficient.

Watch the full interview on YouTube!

Interview Chapters

  • 00:00 Introduction to AI in Industrial Automation
  • 02:52 The Evolution of AI in Manufacturing
  • 05:45 Understanding Different Types of AI
  • 08:27 Generative AI vs. Autonomous AI
  • 11:13 Real-World Applications of Autonomous AI
  • 17:10 Challenges and Considerations in AI Implementation
  • 19:54 Identifying Use Cases for AI in Manufacturing
  • 24:01 The Future of AI in Manufacturing
  • 28:58 Building Optimal Schedules with AI
  • 30:12 Expert Insights and Legacy Knowledge
  • 31:40 Adapting to Equipment Changes
  • 33:31 Aging Infrastructure in Manufacturing
  • 36:06 Competing with Global Manufacturing
  • 38:48 The Need for Automation
  • 41:18 The AI Bubble and Its Implications
  • 43:48 Taking Action Towards AI Readiness

Manufacturing Insights

1. AI is an umbrella term, not a single technology

One of the most important points is that AI in manufacturing isn’t synonymous with ChatGPT. Throughout this podcast episode we describe machine learning, model predictive control, computer vision, autonomous AI, and generative AI as different technologies that fall under the broader AI umbrella. That’s an important distinction for manufacturers evaluating technology.

2. Generative AI isn’t necessarily the best solution for manufacturing problems

This is probably the most provocative part of the podcast episode. Bryan argues that generative AI has significant limitations when it comes to making autonomous decisions on the plant floor, particularly because of hallucinations and its inability to reliably reason through physical cause-and-effect relationships. That doesn’t mean generative AI has no place in manufacturing, but sometimes the flashy technology everyone is talking about isn’t the correct choice.

3. Operational problems require operational AI agents

Bryan draws a particularly useful distinction: “Generative AI is better suited to knowledge problems, while autonomous AI can be suited to operational problems.” A manufacturing problem might be:

  • “What does this machine error mean?” → potentially a knowledge problem
  • “How do I find the procedure for fixing this?” → knowledge problem
  • “How should I adjust this process to get back into specification?” → operational problem
  • “How should the machine respond to changing conditions?” → operational problem

The technology should follow the problem. You don’t want to be using generative AI to solve an operational problem that it wasn’t specifically trained to solve as generative ai is often trained on a wide range of data, which is good for LLMs and Image generation, but it will often hallucinate and give incorrect advice. Not what you want on a factory floor where the stakes are so much higher. AI Agents on the other hand are trained specifically on the data for certain tasks.

4. AI agents can help preserve manufacturing expertise

This is where the conversation connects beautifully to the manufacturing workforce shortage. Manufacturers are losing decades of experience as expert operators reach retirement without any plan to pass on their knowledge. Bryan describes AI agents as a way of capturing some of the problem-solving capability that previously lived inside those experienced employees. They do this by interviewing these expert operators, recording them, capturing data from the machine, and training the AI on all of that by creating a real world simulation for the AI to play around with.

5. AI agents can turn decades of manufacturing experience into a repeatable system

The glass-bottle example is probably the strongest case study in the podcast episode. The manufacturer had a process that would drift out of specification. Only two highly experienced operators were particularly good at bringing the process back into specification, and it could take them 3 to 20 minutes depending on the situation. They used historical process data to built a simulation of the real-world process and trained an AI agent how to return the system to the desired state. This AI agent can then be incorporated into a UI that lets operators input data and received exact directions on how to return to specification.

That’s a fantastic manufacturing AI story because the benefit isn’t abstract anymore. We have a problem that requires an expert → model how the expert solves it → make that capability available and consistent. Essentially turning decades of manufacturing experience into a permanent fixture for all operators to use.

6. AI agents can reduce the manufacturing experience gap

The conversation makes a strong connection to the problems we have been exploring in other Manufacturing Insiders content: manufacturing shouldn’t require decades of experience to operate effectively. Bryan gives the example of a machining problem where an experienced operator might immediately recognize deflection and know which parameters to change. Without that person, a less experienced operator has to guess, experiment, and potentially waste material and time. AI can potentially move some of that expertise into the manufacturing system for the new operator to reference.

7. Adopting AI agents can be harder than building the model

This is another major takeaway. Bryan says manufacturers had already been hiring data scientists who could successfully build models. The problem was getting those models onto the plant floor and into actual operations. That requires much more than data science.

It can involve:

  • Plant-floor integration
  • Controls
  • Networking
  • Data infrastructure
  • Manufacturing expertise
  • Cybersecurity
  • Employee adoption
  • Organizational change

8. Digital transformation requires leadership commitment

Bryan agrees that implementing these systems requires top-down support. If leadership isn’t committed to digital transformation, it becomes very difficult to make the organizational and infrastructure changes necessary to support it. This also connects nicely to the idea that manufacturers shouldn’t view AI as an isolated software purchase. It’s an operational transformation.

9. AI can expose weaknesses in a manufacturer’s existing infrastructure

One of my favorite ideas from the transcript is that AI can become a reason to modernize the factory itself. Bryan suggests that plant-floor employees can use the company’s desire to implement AI as leverage to get leadership to address outdated equipment and infrastructure that prevent the company from accessing the data it needs. In other words: You can’t build a smart factory on top of infrastructure that can’t communicate. That gives you a nice secondary argument around networking, sensors, controls, MES, and data infrastructure.

10. Manufacturing may need AI agents because it can’t simply add more workers

The final section connects AI to global manufacturing competitiveness. Bryan points out that the U.S. manufacturers face a very particular labor environment that is losing expertise, and struggling to cultivate new expertise, and therefore need to investigate automation and AI to increase productivity that is being lost.


All Episodes Become a Guest