In mining maintenance, AI solutions are increasingly available, but not all AI is created equal. Most systems today fall into two categories:
• Traditional (Predictive) AI
• Agentic AI
At a high level, they may appear similar, yet the difference is substantial. Traditional AI provides insights, leaving operational decisions to human teams. Agentic AI, in contrast, delivers context-aware recommendations and outputs that guide human decision-making, significantly improving efficiency, safety, and compliance.
Traditional AI excels at analysing historical and real-time data to predict equipment failures or highlight anomalies. Capabilities typically include:
✔ Predicting potential equipment failures
✔ Identifying anomalies within CMMS data
✔ Tracking performance trends
✔ Displaying dashboards for monitoring
However, Traditional AI is inherently reactive and advisory. Once insights are provided, your team must still interpret the data, plan actions, create work orders, prioritise tasks, verify compliance, and manage labour. In essence, Traditional AI highlights potential issues but does not execute solutions.
Agentic AI extends beyond analytics to assist in decision-making and output generation, acting as a digital reliability engineer. Nexphase exemplifies this approach through a human-in-the-loop system, where the AI guides and informs, but all final decisions and approvals remain with the user.
Key advantages include:
✔ Contextual analysis: Evaluates failure probability, asset criticality, backlog, labour availability, and scheduled shutdowns.
✔ Decision guidance: Recommends optimal maintenance actions based on engineering logic.
✔ Output generation: Prepares fully formed work orders, maintenance strategies, and compliance-aligned documents.
✔ Workflow assistance: Guides users through each step via an interactive sidebar agent, providing answers based on session data.
Traditional AI:
“Your haul truck radiator is trending toward failure in 18 days.”
Agentic AI (Nexphase):
“Based on failure probability, asset criticality, backlog, available labour, and upcoming shutdown:
➡️ The optimal action is to replace the radiator during the shutdown window.
➡️ A draft work order has been prepared with safety steps, required parts, labour hours, and skill requirements.
➡️ Tasks are prioritised with associated cost and risk implications.”
All outputs are generated for review; human oversight ensures safety and operational alignment. Integration with the client’s CMMS for automated execution is available via enterprise agreement and API.
Through guided Agentic AI, Nexphase supports maintenance teams in ways Traditional AI cannot.
Capabilities include:
✔ Extraction of OEM requirements — e.g., “What interval does the radiator coolant get changed?”
✔ MTBF and failure pattern analysis — e.g., “Provide all details for tyre failures on this haul truck.”
✔ Identification of missing maintenance tasks — e.g., “Is changing the hydraulic oil every 250 hours excessive?”
✔ SAE JA1011-compliant RCM logic — e.g., “What drives the recommendation to change the fuel filter earlier than OEM suggests?”
✔ Preparation of fully-formed work orders — ready for review and CMMS integration if applicable
✔ Construction of comprehensive maintenance strategies — completed in minutes with guided steps
✔ Embedding of Worksafe, MSHA, and ISO compliance
✔ Generation of documents ready for CMMS import
Mining operations operate under significant constraints: limited engineering capacity, complex assets, and high consequences for errors.
Traditional AI provides data; Agentic AI provides actionable, context-aware guidance that accelerates decision-making while preserving safety and compliance.
Nexphase was designed from the outset to deliver this capability, enabling teams to:
• Make informed decisions faster
• Reduce manual effort
• Maintain operational and regulatory compliance
• Optimise resource allocation
In short, Nexphase does not replace engineers, it empowers them with intelligent, guided assistance, bridging the gap between insight and execution.