Maximo Mobile and Field Service in MAS 9.2: The Connected Technician

Maximo Mobile and Field Service in MAS 9.2: The Connected Technician

Field service management has always been the sharp end of asset management. It is where plans meet reality, where schedulers in an office try to match technicians in trucks to problems in the field, and where the quality of your data and tools determines whether a job takes two hours or four. IBM Maximo Application Suite 9.2, released in June 2026, brings the most significant set of mobile and field service improvements the platform has seen. AI is now embedded in the technician's daily workflow, scheduling has become conversational, and the gap between the field and the back office is narrower than it has ever been.

This article covers what is new in Maximo Mobile and Field Service Management (FSM) in MAS 9.2, what changed from 9.1, and what implementation teams need to know to deploy these capabilities successfully. We walk through the AI-powered Maximo Assistant on Mobile, conversational scheduling and what-if analysis, Visual Inspection with on-device inference, the expanded offline capabilities, the HSE incident reporter, the centralized mobile administration console, and the deployment architecture that underpins all of it.

Maximo Assistant on Mobile: Natural Language in the Field

The Maximo Assistant on Mobile is the most visible AI feature in the 9.2 release for field technicians. It allows technicians to use natural language to find asset information, review work order history, and complete work efficiently. Instead of navigating through multiple applications and search screens, a technician can type or speak a question: "Show me the maintenance history for pump P-301" or "What were the last three work orders on boiler B-12?" The assistant returns the relevant information in a conversational interface, with links to drill down into specific records.

Under the hood, the assistant uses a retrieval-augmented generation (RAG) architecture. It queries the Maximo Manage database through standard APIs, retrieves the relevant records, and uses a language model to compose a natural language response. The response includes direct links to the underlying Maximo records, so the technician can jump from the assistant's answer to the full work order or asset record with a single tap. The assistant is aware of the technician's security context, which means it only returns information the technician is authorized to see.

The practical impact is significant. Technicians often spend 15 to 20 minutes per job searching for asset history, reviewing previous repairs, and understanding the context of the current work. The assistant compresses that into seconds. For new technicians who do not have years of institutional knowledge about which assets have chronic issues and what repairs have been attempted, the assistant provides instant access to the same information that experienced technicians carry in their heads.

A deployment consideration: the assistant requires connectivity to the Maximo server. It does not work in fully offline mode, because the RAG architecture needs to query the database in real time. If your technicians work in areas with poor connectivity, they should use the assistant when online to gather context before going offline to complete the work. IBM has indicated that offline assistant capabilities are on the roadmap but not available in 9.2.

The assistant also supports voice input, which is a practical feature for technicians wearing gloves or working in environments where typing on a phone is impractical. The voice input uses the device's native speech-to-text capability, and the resulting text is processed the same way as typed input. The response is displayed as text, not spoken back, which is appropriate for noisy industrial environments where audio output would be difficult to hear.

The AI assistant is configured through the AI Configuration application in Maximo Manage. Administrators define which data sources the assistant can query, which user roles have access, and what types of questions it can answer. The configuration also includes guardrails to prevent the assistant from returning sensitive information or making recommendations outside its scope. This is important for organizations in regulated industries where AI outputs need to be controlled and auditable.

Conversational Scheduling and What-If Analysis

On the back-office side, MAS 9.2 introduces AI-enabled conversational scheduling and what-if analysis for planners, schedulers, and field service managers. This feature allows schedulers to explore scenarios using natural language instead of manually adjusting assignments and running calculations.

A scheduler can ask: "What if I add two more technicians next week?" or "What happens if I prioritize the pump overhaul at Plant 3 over the scheduled PMs?" The system applies the organization's constraints and business rules to the query and returns optimized assignment options. It considers technician availability, skill qualifications, travel time, inventory availability, and work order priority. The scheduler can then accept, modify, or reject the proposed changes.

This is a meaningful shift from the manual, spreadsheet-driven scheduling that most field service organizations still rely on. In a typical MAS 8.x environment, a scheduler spends 30 to 40 percent of their day on what-if analysis: manually reassigning work orders, recalculating travel times, checking skill availability, and trying to optimize the schedule. Conversational scheduling compresses this into a few minutes per scenario.

The what-if analysis is not just about adding resources. It can also model the impact of removing resources (sick calls, vacation), changing priorities (emergency work displacing planned work), or shifting schedules (moving a PM from Tuesday to Thursday). The system shows the projected impact on key metrics like first-time fix rate, SLA compliance, and technician utilization.

A practical consideration: the conversational scheduling feature needs clean data to produce useful results. If your technician skill records are incomplete, your travel time estimates are inaccurate, or your work order priorities are inconsistent, the what-if analysis will produce recommendations that look good on paper but fail in practice. Before enabling this feature, invest in cleaning up your labor records, ensuring skill qualifications are current, and validating that your work order priority scheme reflects actual business priorities.

// Example: Conversational scheduling query and response
{
  "query": "What if I add two technicians with MECH-L2 qualification next Tuesday?",
  "context": {
    "date": "2026-08-11",
    "constraints": {
      "skills": ["MECH-L2"],
      "workOrderPriorities": ["1-CRITICAL", "2-HIGH", "3-MEDIUM"],
      "slaComplianceTarget": 0.95
    }
  },
  "response": {
    "projectedImpact": {
      "firstTimeFixRate": "+4.2%",
      "slaCompliance": "+2.8%",
      "technicianUtilization": "78% -> 84%",
      "backlogReduction": "12 work orders"
    },
    "recommendation": "Adding two MECH-L2 technicians on Tuesday would reduce the maintenance backlog by 12 work orders and improve SLA compliance by 2.8%. The main bottleneck shifts from technician availability to parts availability for work orders WO-10472 and WO-10489."
  }
}

Maximo Visual Inspection: AI on the Device

Maximo Visual Inspection in MAS 9.2 introduces AI-based visual inspection with local inference directly on the mobile device. This means a technician can point their phone camera at an asset, and the device runs a trained model to identify defects, wear patterns, or compliance issues without sending the image to a server for processing.

Local inference is a critical capability for field operations. Many industrial sites have poor or no network connectivity in the field. Traditional cloud-based image recognition requires uploading the image and waiting for a response, which is impractical when you are standing next to a compressor in a remote location. With local inference, the model runs on the device, and the result is available in seconds.

The use cases span industries. A technician inspecting a pipeline can point the camera at a weld and the model identifies potential corrosion patterns. An electrician can photograph a breaker panel and the model verifies that the correct breakers are installed. A facilities team can scan a roof and the model identifies damaged tiles. The models are trained using Maximo Visual Inspection's training pipeline, which allows subject matter experts to upload labeled images and train custom models without data science expertise.

A deployment pattern: start with a single inspection type and a small set of assets. Train the model with 200 to 500 labeled images. Deploy the model to a pilot group of technicians. Measure the model's accuracy (true positive rate) and false positive rate in the field. A true positive rate above 80 percent is a good starting point. Below that, add more training images or narrow the scope of the inspection. Once the model is performing well, expand to additional asset classes and inspection types.

The local inference capability uses optimized model formats that are compressed for mobile deployment. The models are trained in the cloud using Maximo Visual Inspection's training pipeline and then deployed to devices through the mobile app's model management system. When a model is updated, the new version is pushed to devices automatically when they are online, ensuring that technicians are always using the latest version. The model sizes are typically 5 to 50 megabytes, depending on the complexity of the classification task, which means they can be downloaded over modest network connections.

Expanded Offline Capabilities and Mobile Improvements

Maximo Mobile has supported offline operations since its introduction, but MAS 9.1 and 9.2 have significantly expanded what technicians can do without a network connection. Understanding the offline model is essential for any organization with field workers in remote locations.

The offline architecture works by preloading data to the device before the technician goes offline. The administrator configures which data is preloaded: work orders assigned to the technician, assets in their work area, spare parts inventory, inspection forms, and relevant job plans. The technician works in the field, creating and updating records offline. When connectivity is restored, the device synchronizes the changes with the server.

In MAS 9.1, the offline capabilities were expanded with several key improvements. Technicians can now reassign work orders even after accepting them, providing flexibility when field conditions change. They can revise accepted work assignments, rejecting or completing them as needed. Meter readings were improved with the ability to add remarks and see the last reading fetched from the server before the meter reading page opens. Follow-up work orders can be created directly from the mobile device, including the ability to choose which assets and locations the follow-up work applies to. Labor hours can be edited and premium labor hours reported directly from the Report Work page.

In MAS 9.2, these capabilities are complemented by the broader AI features. The push notification system has been extended to support background data synchronization, which means the device can sync data even when the mobile app is not actively open. This reduces the time technicians need to wait for synchronization when they come back online, as much of the data transfer happens in the background.

The offline data configuration is managed through Mobile Studio, the low-code application configuration tool. Administrators define which data objects are preloaded, the filter criteria for each object, and the sync frequency. For example, you can configure the app to preload all work orders assigned to the technician for the next 7 days, all assets within a 50-mile radius of their assigned work, and all inspection forms for those assets. This ensures the technician has the data they need without overloading the device storage.

Conflict resolution is an important consideration in offline mode. When a technician edits a record offline that has also been modified on the server, the sync process needs to handle the conflict. Maximo Mobile uses a last-writer-wins approach for most fields, with server-side validation for critical fields. For work order status changes, the server validates that the transition is allowed based on the current server-side status, which prevents a technician from completing a work order that has been cancelled on the server while they were offline.

HSE Incident Reporter and Permit to Work on Mobile

MAS 9.2 introduces the HSE (Health, Safety, and Environment) incident reporter as a mobile capability. 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 HSE incident reporter allows technicians to document safety incidents in real time, including photos, descriptions, severity assessments, and involved parties. The incidents are logged in the Maximo Health, Safety, and Environment application and routed through the appropriate workflow for investigation and corrective action. This real-time capture is a significant improvement over the paper-based incident reporting that many organizations still rely on, where incidents are documented hours or days after they occur and details are lost.

The permit to work process on mobile allows technicians to request, review, and acknowledge permits from their mobile device. This streamlines the permit workflow, which in many organizations still involves paper forms and physical signatures. The mobile permit process supports electronic signatures, ensuring compliance with regulatory requirements while reducing the administrative overhead.

The HSE mobile capability also supports near-miss reporting, which is a proactive safety practice that encourages technicians to report situations that almost resulted in an incident. Near-miss reporting is a leading indicator of safety culture, and making it easy through a mobile app increases the reporting rate, which gives safety teams more data to identify and address hazards before they cause actual incidents.

Centralized Mobile Administration Console

Managing a mobile deployment across hundreds of technicians and multiple sites has historically been a challenge. MAS 9.2 introduces a centralized mobile administration console that gives administrators a single view of their mobile fleet.

The console provides visibility into which devices are active, when they last synced, what data they have preloaded, and what version of the mobile app they are running. Administrators can push configuration changes, deploy new app definitions through Mobile Studio, and monitor sync status across the fleet. This is a significant improvement over the previous approach, where administrators had limited visibility into the state of individual devices.

The console also provides troubleshooting tools. When a technician reports a sync issue, the administrator can check the device's sync history, see which records failed to sync, and identify the cause. This reduces the time to resolve mobile issues and minimizes the impact on field operations.

From an architecture standpoint, the mobile administration console is a web-based application that runs within the MAS environment. It communicates with the mobile devices through push notifications and polling mechanisms. Devices report their status to the console at regular intervals, and the console aggregates the data into a dashboard view. The console does not require direct network access to the mobile devices, which means it works across firewalls and VPN boundaries.

Mobile Architecture and Deployment Considerations

Understanding the mobile architecture is essential for planning a deployment. Maximo Mobile is built as a native application for iOS and Android, available through the Apple App Store and Google Play Store. The app connects to the MAS environment through REST APIs exposed by Maximo Manage, with authentication handled through the MAS Identity and Access Management (IAM) service.

The mobile app uses a local database on the device to store preloaded data and offline changes. The database is encrypted at rest using platform-native encryption (iOS Keychain/Android Keystore). When the app is first launched, the technician enters the MAS server address and authenticates through the standard SSO flow. The app then downloads the configured data sets and remains functional offline.

Deployment planning should consider device management. Most organizations use a Mobile Device Management (MDM) solution to distribute the app, enforce security policies, and manage device inventory. The Maximo Mobile app supports standard MDM configurations including app configuration policies that pre-fill the server address and enforce authentication requirements. This means technicians do not need to manually enter the server URL, reducing support calls and ensuring they connect to the correct environment.

The mobile app is updated independently of the MAS environment. IBM releases mobile app updates through the app stores, and organizations can control the update cadence through their MDM. However, the mobile app version and the MAS version need to be compatible. IBM publishes compatibility matrices for each mobile app version, indicating which MAS versions it supports. Plan your mobile app updates in coordination with your MAS upgrade schedule to avoid compatibility issues.

For organizations migrating from Maximo Anywhere (the legacy mobile platform), the transition to Maximo Mobile requires rebuilding app definitions in Mobile Studio. Maximo Anywhere used a different configuration model based on XML app definitions and the Maximo Integration Framework. Mobile Studio uses a visual designer and a different data model. The migration is manual, but the result is a more maintainable and flexible mobile configuration.

Practical Implications

The mobile and field service improvements in MAS 9.2 are most impactful for organizations with large, distributed field service teams. The AI capabilities (assistant, conversational scheduling, visual inspection) require some investment in configuration and training to deliver value, but once deployed, they fundamentally change how field work is managed and executed.

For smaller organizations or those with simple field service operations, the offline improvements and HSE incident reporter may be the most immediately valuable features. The offline architecture is mature enough to support technicians in truly remote locations, and the HSE capabilities bring safety documentation into the digital workflow where it belongs.

The deployment of Visual Inspection with local inference should be approached iteratively. Start with one use case, prove the value, and expand. The technology is sound, but the model training and validation process takes time. Organizations that try to deploy visual inspection across many asset classes simultaneously tend to struggle with model accuracy and maintenance.

The Bottom Line

MAS 9.2 makes Maximo Mobile a genuinely useful tool for field technicians, not just a data collection app. The AI assistant, conversational scheduling, and local visual inspection are capabilities that would have been science fiction in the 8.x era. They are now production features in 9.2, and organizations that invest in deploying them will see measurable improvements in first-time fix rates, technician productivity, and data quality. The key to success is treating these as operational improvement projects, not just technology deployments. The AI works when the data is clean, the processes are defined, and the technicians are trained.