Workshops

Build a strong foundation in artificial intelligence with this certificate-based course designed for individuals without a deep technical background. This interactive, discussion-based introductory workshop provides a comprehensive overview of AI, making it ideal for beginners. Participants will receive a certificate of completion, professional development hours and a shareable digital badge for social media. It also serves as valuable preparation for the more advanced AI deep dive sessions offered at the conference.

This full-day workshop will cover:

• What AI is and how it works (20-30 minutes)
• Key concepts, including machine learning, neaural networks and generative AI
• Common applications of AI in everyday life and business
• Benefits, limitations and ethical considerations
• Putting AI to use — development and real-world examples

The course blends vendor and end-user perspectives, ensuring attendees leave with a well-rounded understanding of AI fundamentals. Attendees should leave with:

• A stronger understanding of AI concepts
• Practical insights they can apply

Please note: This workshop requires a separate ticket from other conference activities.

Fee: ISA Member: 225 USD; Non-member: 270 USD

Workshop Agenda

09:15-09:45

What is AI and How Does it Work?

This foundational session breaks down the core concepts behind today's artificial intelligence (AI) technologies and provides participants with a shared vocabulary for the rest of the AI primer series. Topics covered include machine learning (ML), generative AI, agentic AI, natural language processing (NLP), computer vision, artificial neural networks (ANN), and convolutional neural networks (CNN). By the end, participants will understand the major branches of AI and how they work together to power the tools shaping our work today.

Speaker: Dave Lafferty, President, Scientific Technical Services

09:45-10:45

AI in Automation Use Cases/Designing AI into Projects

This practitioner-led panel explores how artificial intelligence (AI) can be integrated into automation projects from the earliest design stages through deployment and operations. Drawing on real-world plant and automation examples, end users, system integrators and industrial strategists will demonstrate how AI-assisted engineering tools can help teams design more effectively, accelerate project delivery and improve operational outcomes.

This discussion will feature practical applications of generative AI, natural language processing (NLP), computer vision, agentic AI, convolutional neural networks (CNNs) and artificial neural networks (ANNs), including where each technology fits in real-world industrial environments.

Moderated by Dave Lafferty, this session will use case examples, screenshots, pre-recorded demonstrations and end-user testimony to move beyond concepts and show what AI-enabled automation looks like in practice.

Speakers: 

• Oleg Mikhailov, CEO, Xplorobot

• Brad Davis, President, RIOT Industrial

• Andrew McDonald, Managing Director, FutureofManufacturing, LLC

• Sunil Doddi, Senior Principal Control Systems Engineer, Air Products and Chemicals, Inc.
  
  

11:00-11:45

Lessons Learned from Mimo – A Private LLM Approach

This session will feature an internal case study on ISA’s implementation of its large language model (LLM), Mimo with Betty, offering practical insights into the organization’s approach, lessons learned and key outcomes. Attendees will gain actionable takeaways they can apply within their own organizations, along with a clear “what to do/what not to do” checklist to help guide successful AI adoption.

Speakers:

• Jason Wampler, Managing Director of Information Technology, ISA

• Emily Stamm, Director of Customer Growth, Betty AI
  
  

12:45-13:45

AI for Process Optimization and Autonomous Control

Most plants have automated motion and sequencing. Far fewer have automated judgment. This session walks the three rungs between them: digitizing a line, analyzing what it tells you, and closing the loop so that software holds the process itself.

It runs as a working session rather than a lecture. Three times we will ask the room where it actually stands, on data, on analytics, and on closed-loop control, and each round ends with a live screen rather than a slide.

Along the way we will define the terms that usually go unexplained: what it means to learn the physics of a line from its own sensor data, what reinforcement learning is and how it differs from the predictive maintenance you have already met, and where each of them stops.

We will show results from rubber and thermoplastic extrusion across eight plants, along with the guardrails, deployment phasing and OT security questions that make autonomous control acceptable on a real line. You will leave with a scoring worksheet for your own process and a first step that costs nothing.

Speakers:

• Joseph Hernandez, COO & VP of Engineering, Liveline Technologies

• Sean Scott, VP of Solutions, Liveline Technologies
  
  
• Cissy Gu, Senior Data Scientists, Liveline Technologies

13:45-14:45

AI Agents in Industrial Settings

Large language models (LLMs) have quickly become part of everyday life, but there is a significant difference between an AI system that responds to a prompt and one that can reason through a task, use tools, interact with other systems and act autonomously. This session will explore what makes an AI system an "agent," how agents are designed and built and how those capabilities can be applied in industrial environments. We will look beyond the terminology to examine the mechanisms behind agentic AI and the integration points needed to connect agents with industrial data, software, workflows and existing systems. We will then focus on where agents can provide practical value in process manufacturing and other industrial settings today, as well as where caution is required.

Attendees will explore potential use cases, deployment approaches and the role of human oversight, while also examining challenges such as reliability, cybersecurity, data quality and safety. The goal of this workshop will be to provide a balanced view of the opportunities and limitations of agentic AI, helping participants understand where agents can be useful and where they can be deployed, where they may not yet be ready for deployment and how to begin evaluating the technology responsibly.

Participants in this session should expect to come away with answers to the following questions:

• What makes up an AI agent? The core components and capabilities that distinguish an agent from a traditional LLM application.

• How do you build an agent? The tools, integrations, workflows and controls required to create and deploy one.

• Where can agents be used today? Practical industrial use cases where agents can provide meaningful value.

• What are the limitations and risks? The technical, operational and security challenges you need to consider before deploying an agent.

• How do you get started? How to identify appropriate first use cases and begin experimenting with agentic AI in a controlled and practical way.

Speaker: Forrest Shriver, CEO, Sentinel Devices

15:00-16:00

Responsible AI Implementation in Manufacturing: Ethics and Governance

Artificial intelligence (AI) is moving rapidly into manufacturing, from machine learning for predictive maintenance and computer vision on the factory floor to generative AI for engineering, purchasing, documentation and customer-facing work. But responsible implementation begins with a deceptively simple question: where does AI actually belong?
AI has not invented bad automation. Manufacturers have long had to contend with poorly designed processes, misleading models and misplaced trust in data and systems that are given more authority than they should have. AI can make those mistakes easier to scale—and harder to recognize—because increasingly capable systems can produce answers, recommendations, software and actions that appear remarkably convincing even when they are wrong.

This session examines those challenges through concrete examples drawn from manufacturing, automation and human-AI interaction. What can we learn from a manufacturer whose purchasing and accounting models accurately reflected its costs yet led management to the disastrous conclusion that it should eliminate internal manufacturing altogether? Why do safety-critical systems sometimes use independent, redundant channels that shut down a process when they disagree? When is computer vision an excellent way to determine whether an area is clear, and when should a conventional light screen, force sensor, interlock or other deterministic safeguard remain the final authority? And what changes when an AI assistant moves from helping an engineer write or analyze PLC code to being allowed to alter the state of an automation cell itself?

The same questions arise away from the factory floor. Generative AI can help an individual employee draft a document, analyze purchasing data, investigate lead times or communicate with a customer much faster. But organizations should distinguish between augmenting an employee and automating an entire business process. Sometimes the better solution is not AI at all, but conventional process engineering: a well-designed database, dashboard, workflow or purchasing tracker may be more reliable, transparent and maintainable than introducing a probabilistic system simply because AI is available.

These examples point to a broader issue at the heart of AI ethics and governance: authority. There is an important difference between a system that informs, one that recommends and one that acts. The consequences of an incorrect prediction about future maintenance needs are very different from those of an incorrect purchase order, customer commitment, PLC modification or safety decision. Responsible AI implementation, therefore, requires organizations to consider not only whether an AI system performs well but also what happens when it fails, whether that failure is detectable and reversible, and what safeguards should exist between a model's output and consequential action.

Drawing on research in human-AI collaboration, experience with small manufacturing organizations, and examples from industrial automation, this talk will explore how manufacturers can leverage increasingly capable AI without abandoning the engineering principles that made modern automation reliable. The future of manufacturing is unlikely to be a choice between people, conventional automation and AI. The more interesting—and more important—challenge is deciding what each should be allowed to do.

Speaker: Johnathan Mell, Assistant Professor, ScionAI Lab, University of Central Florida