AI Inside Maximo in 2026: From Predictive Maintenance to Agentic Condition Insight

AI inside Maximo is no longer a research project. As of the Maximo Application Suite 9.2 release in June 2026, every reliability engineer, planner, technician, and safety lead in a Maximo shop is working alongside AI capabilities that have shipped to production in hundreds of organizations. The shift from standalone AI modules to workflow-native AI is the defining change in this release, and it changes how organizations should plan their adoption path.

MAS 9.2 expands AI across reliability insights, field execution, safety and compliance workflows, document-based information extraction, and orchestration across systems. It introduces agentic workflows designed to help guide decisions and move work forward across these processes in practical, operationally grounded ways. As IBM stated in its announcement: "IBM is announcing Maximo Application Suite 9.2, a release that brings asset-first AI directly into the workflows that reliability, maintenance, field service, safety and operations teams use every day." The emphasis is on operationalizing AI, not experimenting with it.

This article covers the five AI capabilities in MAS 9.2, the watsonx platform that powers them, the maintenance maturity model that contextualizes them, and a phased 12-month adoption roadmap for organizations starting their Maximo AI journey.

The watsonx Foundation

The AI capabilities in MAS 9.2 are built on IBM watsonx, IBM's AI and data platform. watsonx provides the machine learning, natural language processing, and generative AI capabilities that power Condition Insight, Visual Inspection, Maximo Assistant, and the new agentic workflows. The platform consists of three components: watsonx.ai for model training and serving, watsonx.data for data ingestion and preparation, and watsonx.governance for model lifecycle management and compliance.

For Maximo deployments, this means the AI models run on the same infrastructure as Maximo, with the same security, governance, and compliance controls. Data does not leave the Maximo environment to reach an external AI service. The models are trained, deployed, and monitored within the watsonx platform, and Maximo calls them through authenticated, audited API calls. This architecture is critical for organizations in regulated industries where data residency and model auditability are compliance requirements.

watsonx provides model governance capabilities that ensure AI models are traceable, versioned, and auditable. Every model has a documented lineage from training data through deployment, and every prediction or recommendation can be traced back to the specific model version that produced it. This governance framework is essential for organizations that need to explain why an AI system recommended a particular maintenance action, especially in safety-critical environments. The governance lifecycle includes use case submission, risk assessment, model testing and validation, production infrastructure setup, and ongoing monitoring with threshold breach detection. Each stage has defined roles, approval workflows, and documentation requirements.

For Maximo Predict, the training data (work order history, failure records, meter readings, sensor data) flows from Maximo through watsonx's data ingestion pipeline into the model training environment. The data scientist works within watsonx.ai to build and train models using Jupyter notebooks, selecting features, tuning hyperparameters, and evaluating model performance against held-out test data. Once a model meets the acceptance criteria, it is deployed to watsonx.ai's serving infrastructure, where it generates predictions on demand or on a scheduled basis.

The model behind Maximo Assistant has evolved rapidly. As of November 2025, IBM deprecated the Granite 3.2 8B Instruct model in favor of the gpt-oss-120b model, which provides significantly better accuracy on complex queries. The assistant uses the nl2oslc model template, which translates natural language questions into OSLC REST API queries against the Maximo data model. This means the assistant can access any object in the system, including custom objects, and supports a wide range of queries without requiring custom configuration for each new object type.

The Five AI Capabilities in MAS 9.2

Maximo Condition Insight

The single most important AI capability in MAS 9.x is Maximo Condition Insight, which IBM introduced in late 2025 and which has matured into a flagship feature by mid-2026. Condition Insight is an agentic AI capability within Maximo Asset Performance Management (APM) that interprets asset data to explain asset condition, highlight emerging trends, and recommend corrective actions. It works in concert with the other MAS applications to deliver a unified, condition-based maintenance approach across the Maximo ecosystem.

The technical workflow is as follows. Condition Insight continuously monitors the assets that have been configured for condition-based maintenance. When it detects a change in asset behavior, such as a rising vibration trend, an increase in work order frequency, or a meter reading outside normal range, it triggers an evaluation. The evaluation aggregates all available data for the asset, including recent work orders, inspection results, meter history, FMEA records, and real-time sensor data from Maximo Monitor. The AI agent reasons about this data, identifies the pattern, and produces a plain-language summary of the asset's condition, the emerging trend, and a recommended action.

The recommendations are not generic. They are specific to the asset, its operating context, and its failure history. For example, Condition Insight might recommend: "Pump P-104 shows rising vibration over the last 30 days, with work order frequency increasing 40% compared to the prior quarter. FMEA record FMEA-PUMP-003 identifies bearing wear as the likely failure mode. Recommend scheduling a vibration analysis within 7 days and preparing bearing replacement kit BK-2274."

The agentic nature of Condition Insight means it does not simply produce a score or a probability. It reasons about the available data, explains its reasoning in natural language, and suggests a specific next action. This is a fundamental shift from the predictive model approach, which outputs a failure probability and leaves the interpretation to the reliability engineer. Condition Insight closes the gap between data and action, which is where most predictive maintenance programs stall.

Maximo Predict

Maximo Predict is the application that started the AI journey in Maximo, and it remains the foundation for everything MAS 9.2 builds on. Predict uses AI and machine learning to analyze historical maintenance records, operational data, inspection reports, and environmental data to predict downtime, degradation, and failures. The June 2026 MAS 9.2 release also decouples the Predict component from the IoT service layer. Predict now operates independently, with IoT-related functionality accessed through Monitor or the appropriate integration layer. This improves modularity and simplifies the deployment architecture for customers running predictive models without a full IoT footprint.

The implementation workflow for Predict is well-documented and follows a clear sequence:

1. Create asset groups based on asset type, criticality, or operational context. Predict works best when analyzing groups of similar assets rather than individual assets in isolation. A group might be "all centrifugal pumps at the north plant" or "all transformers rated above 500 kVA."
2. Train predictive models using the default notebooks provided with Predict. A data scientist works with the group ID and uses Jupyter notebooks to build and train model instances. The notebooks handle the common algorithms, and the data scientist focuses on feature engineering and model selection.
3. Deploy the trained model so that it generates predictions for each asset in the group. Predictions include current failure probability and estimated failure date.
4. Monitor predictions through the Predictions section in Maximo, which shows failure probability, estimated failure date, and recommended actions for each asset.
5. Use work queues to track assets with high failure probability or assets predicted to fail before the next scheduled PM work order. These work queues become the daily planning tool for reliability engineers.

An Asia Pacific oil and gas producer that implemented Maximo Predict achieved 87% predicted failure accuracy, with some models consistently providing 100% accurate results. This proactive approach avoided $10 million in missed revenue by preventing unplanned critical failures, increased production rates, and improved maintenance and replacement strategies. The key to this success was sufficient historical data: the producer had years of work order history, failure records, and sensor data for the critical assets that Predict was trained on.

Maximo Monitor

Maximo Monitor (formerly Maximo Health) provides AI-enabled remote monitoring of asset condition at scale. It ingests sensor data from IoT devices, calculates asset health scores, and triggers alerts when condition indicators move outside normal range. Monitor is the data ingestion layer that feeds Condition Insight and Predict, but it also provides standalone value through its dashboard and alerting capabilities.

Monitor supports a wide range of industrial protocols through managed gateways, including Modbus, OPC UA, MQTT, and HTTP-based REST APIs. The managed gateway architecture allows organizations to connect to industrial control systems without exposing them directly to the internet, which is critical for operational technology security. Data flows from sensors through the managed gateway into Monitor, where it is processed, stored, and made available to the other MAS applications.

The asset health score in Monitor is calculated using a combination of sensor thresholds, historical patterns, and AI models. The score provides a single, normalized indicator of asset condition that maintenance teams can use to prioritize work. When the score drops below a configured threshold, Monitor generates an alert that can trigger a work order, notify a reliability engineer, or feed into Condition Insight for a more detailed analysis.

Maximo Visual Inspection

Maximo Visual Inspection uses computer vision models to detect defects from images and video, with the option to run inference locally on mobile devices. This capability does not depend on the historical data that Predict requires, so it can be deployed in parallel with predictive models.

The implementation process for Visual Inspection involves building a labeled image library for the most common inspection types, training computer vision models on those images, and deploying the models to mobile devices for field use. Technicians take photos of equipment during inspections, and the model identifies defects in real time, flagging issues that might be missed by visual inspection alone.

The mobile inference capability is particularly important for field operations in remote environments. Running the model on the device eliminates the latency of sending images to a cloud endpoint for processing, and it allows the inspection to proceed even without network connectivity. The model results are stored locally and synchronized with Maximo when connectivity is restored.

Visual Inspection is particularly valuable for inspections that are currently manual, time-consuming, and subject to human variability. For example, inspecting corrosion on pipeline segments, identifying cracks in concrete structures, or detecting wear on conveyor belts can all be automated with computer vision models that apply consistent criteria across every inspection.

Maximo Assistant on Mobile

Maximo Assistant on Mobile helps technicians use natural language to find asset information, review history, and complete work efficiently in the field. Instead of navigating through application menus and running reports, technicians can ask questions and get answers. The assistant can access any object in the system, including custom objects, and supports a wide range of queries.

Example interactions include: "Show me the work order history for pump P-104", "What are the open work orders for location BUILDING-A?", "Create a work order to replace the bearing on motor M-203", and "What safety permits are active for asset TRANSFORMER-12?" The assistant translates these natural language requests into OSLC REST API calls, retrieves the data, and presents it in a readable format.

The assistant benefits from having a broad user base, so rollout should target the entire technician population rather than a small pilot group. It does not require a strong data foundation, which makes it a good starting point for organizations that want to build user familiarity with AI capabilities before deploying the more advanced features that require clean historical data.

The Maintenance Maturity Model

The AI capabilities in MAS 9.2 map to a maintenance maturity model that helps organizations understand where they are and where they are heading:

• Stage 1: Reactive - Fix when it breaks. No AI needed.
• Stage 2: Preventive - Fix on a schedule. No AI needed, but Maximo Manage provides the scheduling framework.
• Stage 3: Condition-based - Fix when condition indicators trigger. Maximo Monitor and Health provide the condition monitoring and scoring.
• Stage 4: Predictive - Fix before it fails. Maximo Predict provides the failure forecasting.
• Stage 5: Prescriptive - AI recommends the optimal action. MAS 9.2 Condition Insight provides the agentic recommendations.

Most organizations in capital-intensive industries are at Stage 2 or Stage 3. The goal of a Maximo AI deployment is to move the organization to Stage 4 and eventually Stage 5, where AI not only predicts failures but recommends specific corrective actions with reasoning that maintenance teams can evaluate and act on. The transition between stages is not automatic. Each stage requires specific data, specific capabilities, and specific organizational changes. Moving from Stage 2 to Stage 3 requires sensor data and condition monitoring. Moving from Stage 3 to Stage 4 requires historical failure data and trained predictive models. Moving from Stage 4 to Stage 5 requires the agentic AI capabilities in Condition Insight and the organizational willingness to let AI recommend actions, not just predict failures.

Safety and Compliance: AI-Assisted Incident Classification

For organizations focused on safety and compliance, MAS 9.2 connects these responsibilities more directly to daily operations. Expanded capabilities across asset-centric waste management, contractor safety oversight, and mobile-first safety workflows make it easier to manage compliance, track risk, and act in real time.

AI-assisted incident classification helps improve the consistency and completeness of reporting by suggesting categories and identifying similar events. When a technician records an incident on a mobile device, the AI suggests the most likely incident category based on the description and the asset involved. It also identifies similar incidents that have been recorded, which helps detect patterns that a single incident might not reveal. This pattern detection is particularly valuable for identifying systemic issues that contribute to repeat incidents across different locations or asset types.

For example, if a technician records a "slip hazard on platform" incident at Site A, the AI might identify three similar incidents at Sites B and C over the past six months, all involving the same type of platform surface material. This pattern would not be obvious from any single incident report, but the AI's ability to search across all incident records and identify similarities makes it visible. The safety team can then prioritize a systemic fix (replacing the platform surface material) rather than treating each incident as an isolated event.

Field teams can capture incidents, complete inspections, and initiate permit-to-work processes directly on a mobile device, improving data accuracy and responsiveness while reducing reliance on manual or disconnected processes. The safety capabilities are most effective when safety is part of daily operations rather than treated as a separate process that happens after the maintenance work is done.

Practical Implications

The recommended sequence for organizations starting their Maximo AI journey in 2026 is a phased approach that builds the data foundation while delivering incremental value at each step.

Month 1 to 2: Audit the data foundation. Look at the asset hierarchy, the meter history, the work order history, and the failure records. Identify the asset classes where the data is complete and consistent enough to support AI. This audit will reveal that 20 to 30 percent of assets have data quality good enough for Predict, while the rest need remediation. Document the gaps and prioritize remediation for the most critical asset classes.

Month 3 to 4: Deploy Maximo Assistant on Mobile. This does not require a strong data foundation and provides immediate value to field technicians. It also builds user familiarity with AI capabilities in the Maximo context, which helps with adoption of the more advanced features later.

Month 5 to 6: Pilot Maximo Predict on one asset class. Choose the asset class with the best data quality from the audit. Train the default models and deploy them. Measure the prediction accuracy and the impact on maintenance planning. This pilot will demonstrate the value of predictive maintenance to executive sponsors and justify further investment.

Month 7 to 8: Deploy Maximo Visual Inspection. Choose two or three high-value inspection types where computer vision can augment human inspection. Train models and deploy them to mobile devices. This does not depend on the historical data that Predict requires, so it can run in parallel.

Month 9 to 12: Pilot Condition Insight. Deploy Condition Insight on the same asset classes where Predict and Monitor are already running. The agentic recommendations layer on top of the underlying predictions, so the data foundation from the Predict pilot supports the Condition Insight pilot. Measure the impact on maintenance decision-making and the quality of the AI-generated recommendations.

Month 12 and beyond: Scale and mature. Expand to additional asset classes, integrate additional data sources, and begin building agentic workflows that coordinate across multiple systems. The MCP Server in MAS 9.2 enables these workflows by allowing AI agents to interact with Maximo data programmatically. Start with the simplest agentic workflows (safety incident triage, MOC routing) and build up to more complex orchestration.

Bottom Line

AI inside Maximo in 2026 is asset-first, workflow-native, and grounded in the watsonx platform. The five capabilities (Condition Insight, Predict, Monitor, Visual Inspection, and Mobile Assistant) are not standalone tools. They are integrated layers that build on each other, and the implementation sequence matters. Start with the data audit, deploy the Mobile Assistant for immediate value, pilot Predict on your best-data asset class, add Visual Inspection in parallel, and build up to Condition Insight. The maintenance maturity model provides the roadmap: move from reactive to preventive to condition-based to predictive to prescriptive, one stage at a time, with measurable outcomes at each step.