How Much Prep Work Do Trusted AI Agents Really Need on the Factory Floor?

Deploying trusted AI agents in manufacturing requires the majority of effort in data integration, context modeling, and governance before any model runs. According to the “Agentic AI in Manufacturing: From Copilots to Autonomous Systems” panel, sponsored by Cybus and Infinite Uptime, operators need agents that surface timely, actionable recommendations inside their existing HMI, SCADA, and MES tools, starting in advisory mode before autonomy is granted.

Below is a practical synthesis of the panel’s answer, focused on what manufacturers can actually do next.

What prep work do AI agents need in manufacturing?

Preparing a manufacturing environment for AI agents requires building a data foundation first. The “Agentic AI in Manufacturing” panel found that the majority of effort falls on data integration, metadata definition, governance, and operator change management. Unified, low-latency access to real-time and historical data across HMI, SCADA, MES, and PLM systems is the starting point.

• Unified access to the essentials. Agents need low-latency, real-time streams for detection and response, plus historical data to learn patterns. This unified access doesn’t happen by accident: It requires a central, secure and scalable data foundation. Such a digital backbone enables seamless connectivity across HMIs, SCADA, MES, CMMS and PLM systems. 
• Context beats volume. Metadata—units, ranges, states, equipment relationships—was called the most underestimated barrier. Without clear meaning, agents can’t reason about thresholds, severity, or trade-offs.
• Governance and ownership. Define who owns which datasets and decisions, ensure auditability, and build an agile stack that can onboard new tools quickly. The organizational setup matters as much as the technology.
• Edge-plus-cloud by design. Keep time-critical detection and actions near the line (edge) and use cloud for aggregation, training, and coordination. Control cost and complexity by reusing common data models and connectors across use cases.
• Standardization and interoperability. Avoid bespoke adapters plant by plant. Standardize sensing and data interfaces early so new lines don’t become one-off engineering projects.
• People and change. Operators must see agents as partners. That takes transparency, training, and staged rollouts that prove value in their day-to-day tools.

The session didn’t offer a percentage split of effort, but the message was clear: expect the majority of work to be in data plumbing, context modeling, integration, and governance—not in the last mile of model tuning.

When should AI agents act without human approval?

AI agents in manufacturing should begin in advisory mode and earn autonomy gradually. The panel recommended starting with human-in-the-loop validation at the HMI and SCADA level, then expanding autonomous action at the MES layer (ISA-95 Level 3) for well-understood, auditable tasks. Safety-critical and regulated steps should remain gated by expert review, with explainability and rollback paths maintained at every stage.

• Early phase: advisory. Surface recommendations inside the tools people already use (HMI/SCADA/MES). Humans validate and act, building trust while capturing feedback for improvement.
• Respect ISA-95 layers. At Levels 0–1 (sensors, actuators, controllers), agents should monitor and advise—not move setpoints. Safety and process integrity come first.
• Where autonomy grows first. Level 3 (MES) is the “sweet spot” for orchestrating workflows, scheduling, and preparing work orders. As validation, reliability, and audit trails mature, you can reduce mandatory expert review for specific, well-understood actions.
• Guardrails never go out of style. Maintain explainability and a safe rollback path, especially in regulated settings.

How Can Manufacturers Deploy AI Agents This Quarter?

Manufacturers can begin deploying AI agents in a structured way within a single quarter by following a focused, six-step readiness sequence. The approach starts with a single outcome and asset family, standardizes only the signals actually needed, and stages autonomy from advisory recommendations through to auto-approval for narrowly scoped actions once performance is proven across shifts

What are the risks of AI agents on the factory floor?

Risk design decisions must happen before AI agents go live in manufacturing. Safety and security at PLC and controller interfaces need network segmentation to block uncontrolled setpoint changes. Regulated environments require compliance and explainability built in from day one. Redundant per-use-case infrastructure drives cost creep, so the panel recommended reusing shared data models, connectors, and services across use cases.

What results should AI agents deliver in manufacturing?

After a successful AI agent deployment in manufacturing, similar assets across lines and sites should present consistent data, with onboarding measured in weeks rather than quarters. Operators trust recommendations because they are timely and embedded in daily tools. Finance can track movement on operational KPIs across all shifts, regardless of which teams are on duty.

Takeaway for manufacturers

If you want agents that act with confidence at midnight, invest in the daylight work: consistent signals, meaningful context, clear governance, and operator-first rollouts. Begin with advisory recommendations, prove value where it’s safest and most visible, and expand autonomy as reliability and auditability mature. Expert vetting is essential at first and in safety-critical steps—but it doesn’t have to be permanent.

Source: “Agentic AI in Manufacturing: From Copilots to Autonomous Systems”, sponsored by Cybus and Infinite Uptime, Inc., part of the AI Frontiers 2025 IIoT World’s event.

Related articles:

• A Complete Maintenance Team Made of AI Agents — and Why It Changes Everything
• The Rise of AI Agents: Redefining Automation in Manufacturing