Agentic leaps past Gen-AI in its ability to solve production plant problems

What you’ll learn:

Gen-AI is able to create new content based on the models it has been trained on, but it doesn’t inherently solve problems.
Agentic AI is designed to proactively identify and solve problems and can autonomously detect anomalies and provide solutions.
Agentic is particularly beneficial in manufacturing, where identifying and resolving issues promptly can significantly impact productivity and efficiency.
Critical thinking, coordination, and oversight by humans is essential in using agentic AI. Human workers must validate the outcomes.

Editor's note: Tim Gaus is a semi-regular contributor to Smart Industry. Most notably he was our guest about a year ago for a popular webinar, "Getting the ‘Ts’ to get along."

Manufacturers have historically been early adopters of various AI technologies, starting from classic AI and machine learning to predictive analytics and deep learning.

Over the past year, generative AI has gained significant traction in the industry because of its potential to reduce dependencies on legacy knowledge sources and help workers find answers faster than they’ve been able to before.

Agentic AI marks the beginning of a new phase for Gen-AI, increasing its level of responsibility and autonomy to create a factory management solution of the future.

Gen-AI is distinct in its ability to create new content based on the models it has been trained on. However, it does not inherently solve problems. This is where agentic AI comes into play. AI agents are reasoning engines that can understand context, plan workflows, connect to external tools and data, and execute actions to achieve a defined goal.

Unlike Gen-AI, however, agentic is designed to proactively identify and solve problems and can autonomously detect anomalies and provide solutions. This makes agentic AI a powerful tool that manufacturers should consider implementing in their operations.

AI’s next evolution

Fundamental differences between Gen-AI and agentic AI lie in their core functions—and agentic AI could be the evolution that moves the needle on deployment. While Gen-AI focuses on content creation, agentic is geared towards problem-solving.

What is your company doing about cybersecurity?

For example, Gen-AI might provide a summary of potential issues (e.g., this could be why a machine broke down and here are several ways to fix it), whereas agentic AI can pinpoint a specific problem and suggest solutions (e.g., this is why the machine broke down and here’s the exact part you need to resolve it).

This capability is particularly beneficial in manufacturing, where identifying and resolving issues promptly can significantly impact productivity and efficiency.

Webinar ahead: Great Questions About Tariffs & Global Trade: Planning Amidst Uncertainty

Manufacturers who have already started on their Gen-AI journey can leverage agentic AI to build on these investments. By applying agentic AI to existing data, manufacturers can enhance their machine learning models and improve decision-making processes.

Agentic AI can act as a digital continuous improvement engineer, providing actionable insights without the need for extensive manual analysis. Both individual AI agents with specialized tasks, and multi-agent systems that share task knowledge have their place in manufacturing.

Improving human functions

The introduction of agentic AI could also help bridge existing skills gaps—a critical challenge that leaders have been testing solutions for. By making the workforce more effective, agentic AI holds the potential to keep plants productive even amid an industry-wide talent crunch.

For example, consider a team that doesn’t have much legacy knowledge on troubleshooting a maintenance problem.

Instead of searching for hours through manuals or spreadsheets and then testing several solutions, agentic AI could help a continuous improvement engineer receive direct prompts on what issues to tackle, while a supervision manager can get a detailed report on what happened during the last shift and the recommended course of action.

This proactive approach to troubleshooting, automatic detection of anomalies, and faster problem-solving can significantly enhance workforce efficiency.

However, agentic AI is simply a tool and critical thinking, coordination, and oversight by humans is essential in using the technology. Human workers must validate the outcomes and actions generated by agentic AI, so it’s important that organizations upskill their workforce and teach the technical skills needed to work alongside these tools.

Overall, fostering a workforce that embraces both technology and critical thinking will be one of the most important factors for success. When humans understand new tools and are properly trained with them, they’ll trust them. Without trust, they will not utilize the technology effectively.

Once humans embrace working alongside AI agents, they’ll get to work on the critical problems they haven’t had time for in the past, ultimately allowing them to solve bigger problems and find new pockets of value for the organization.

Key considerations

Although enterprise leaders are leaning into the benefits and potential of agentic AI, adoption at the shop floor level could be challenging due to the rapid pace of change and skepticism from frontline workers. Leveling the field will mean committing to a tech-driven culture and providing support structure inside and outside the plant.

Data quality also remains a significant barrier since only high-quality, accurate inputs can create meaningful outputs. Manufacturers with a strong data foundation are better positioned to leverage agentic AI, while those without it will need to tackle substantial foundational work before they can really leverage this tech.

Episode 2 of (R)Evolutionizing Manufacturing: Data is everything

Manufacturers should avoid implementing technology for its own sake. The focus should be on achieving tangible outcomes and proving ROI—and that’s where balancing foundational investments is essential. After that leaders can target specific pockets of their business with specialized, single-agent workflows.

By starting small and scaling fast in niche functions, leaders get a clear view of the impact and workers have a clear North star, avoiding muddled metrics and overextensions of teams.

From there, teams can explore multi-agent workflows, employing multiple, role-specific AI agents to understand requests, plan workflows, coordinate role-specific agents, streamline actions, and collaborate with humans to validate outputs.

The future of agentic AI in manufacturing is promising, offering a transformative opportunity for the industry that aligns with core lean Six Sigma processes and provides a huge step toward self-healing manufacturing systems. It could be the tipping point for GenAI adoption.

Podcast: AI's role in Kaizen, lean, technology and continuous improvement

As manufacturing operations become smarter and software-defined, having an AI agent that can not only monitor inventory and identify machine issues before they occur, but also automatically negotiate with suppliers, place material orders, and reroute workflows to other machinery as needed to keep production on course—could prove too valuable to ignore.

By understanding its potential and implementing it effectively, manufacturing leaders can drive significant improvements in productivity, efficiency, and overall business outcomes.