All learning series · Suite application · Part 1
Maximo Visual Inspection
Follow images from labeling to evaluation and deployed inference.
Start here
Confirm this product is ready, then use the lesson. The readiness checks are the gate for configuration, for a first journey and for production.
- Are you ready to use Maximo Visual Inspection? Licensing and AppPoints, the records that must exist, and three checklists.
- Basics
- Configuration
- Practical example
- Practice and verify
- Advanced concepts
Are you ready to use Maximo Visual Inspection?
Visual Inspection trains a model on images you have labelled. The 9.2 guidance prices the application at Premium and the install at 45 AppPoints. Each edge inspection is a further install point.
Before configuration
Have the Premium user, the 45-point install, and a set of images you are allowed to use. Edge devices are a second decision.
Before the first successful journey
One question, two labels, and enough labelled images to train and to hold some back.
Before production use
The label guide and the camera install have owners. An edge device is 1 AppPoint per installed inspection in the guidance, and a camera is the usual image source.
Licensing, entitlement and AppPoints
The MAS 9.2 licensing guidance prices Visual Inspection at the Premium user tier and requires a per-install licence of 45 AppPoints. Visual Inspection Edge is a further install: 1 AppPoint per installed inspection. The guidance notes that a camera is typically the image source, so one installed inspection is often one camera. Those install points are not the Premium user points. Predict is Premium without this install line. Images from a site you do not have rights to use are not a training set.
Keep four facts separate for Maximo Visual Inspection. Entitlement is the proof of entitlement and the License Information for IBM Maximo Application Suite. The MAS 9.2 licensing guidance supplements that agreement. Available AppPoints are the pool shown on the AppPoints Usage dashboard: Suite administration, then Suite, then Administration, then AppPoints Usage. Assigned access is the security group together with the user type on the person: Limited, Base or Premium, and Authorized or Concurrent. Consumption is what that assignment draws. Authorized users and administrators reserve AppPoints continuously. Concurrent users draw AppPoints while they are signed in. Install AppPoints stay reserved for the life of that environment. A non-production install of a component draws no install AppPoints when the same component is already installed in production. Administrators and User Management users open the dashboard. Other people need the LICENSEUSAGE security group. Compliance enforcement and idle timeout sit on the dashboard Configuration tab and are suite-administrator settings.
The same guidance publishes these user conversion ratios: Limited is 5 AppPoints for one concurrent user and 2 for one authorized user; Base is 10 concurrent and 3 authorized; Premium is 15 concurrent and 5 authorized; an administrative Base user is 10 AppPoints and an administrative Premium user is 15, both as authorized users. A Base administrator who already reserves 10 AppPoints can be given Premium application access with no further AppPoints. Those ratios were read from the MAS 9.2 licensing guidance on 18 October 2026. Your agreement sets how many AppPoints you purchased. An application that opens shows that the component is deployed. Read the proof of entitlement and the dashboard as well.
Installed and enabled applications
Confirm Visual Inspection is deployed and the 45-point install shows on the dashboard. An edge device, if you use one, shows its own 1-point install. Installed means the application opens. Configured means a dataset and labels exist. Usable means a held-out image returns a label you can compare with the guide.
Foundation configuration
The asset or location in the picture is a Manage record when the result must become work. The images, the labels and the question are the inspection product’s own data. No meter, calendar or map is required for the first model.
Product configuration
Write the question and the label rules. Label the training images. Hold some images back. Grant labellers and model publishers different groups. Review a default confidence threshold before the label is written onto a work order.
Infrastructure and integration
Training runs in the Visual Inspection deployment. An edge device needs a route to receive the model and to return a result, plus power and a mounting location. Storage has to hold the images. The 45-point and 1-point installs do not size the disk. Size it for the images you will keep.
Data and operational readiness
The first journey needs a consistent label on the training set and one held-out image. The result is a label and a confidence, compared with the guide. A missing label is data. A missing install line is entitlement. A result on the training image only is not the test.
Setup order for an inspection model
- 1. Premium user and the 45-point install Both. Edge is 1 point per installed inspection, later.
- 2. A question and label rules What the picture is being asked.
- 3. Labelled images and a holdout The holdout was not used to train.
- 4. A result compared with the guide Then, if you deploy an edge camera, its own install point.
| Requirement | Why needed | Mandatory/Conditional/Optional | Where to configure or verify | Minimum working example | Verification procedure | Consequence if missing | Remediation | Source |
|---|---|---|---|---|---|---|---|---|
| Premium user entitlement | Visual Inspection is a Premium application. | Mandatory | Suite administration → Suite → Administration → AppPoints Usage. Manage → Security → Security Groups, and the suite user type. | A model publisher on Premium. | Confirm Premium on the dashboard for that user. | A Base user is given the application because they can open Health. | Use Premium. | Licensing in Maximo Application Suite 9.2 |
| Install of 45 AppPoints | The guidance allocates 45 AppPoints per Visual Inspection install. | Mandatory | Suite administration → Suite → Administration → AppPoints Usage, install line. | One install. Non-production draws no install points when production already has Visual Inspection. | Match the dashboard to 45. | The application is deployed and the install line is missing. | Apply the 45-point install. | Licensing in Maximo Application Suite 9.2 |
| Edge install of 1 AppPoint per inspection, when you deploy a device | The guidance prices each installed edge inspection at 1 AppPoint and treats a camera as the typical source. | Conditional | The edge device record and the install line. | One camera, one edge install, after the core model works. | Count installed inspections and compare with the 1-point lines. | Cameras are deployed and only the 45-point install is recorded. | Add the edge install for each inspection, or leave the cameras out of this phase. | Licensing in Maximo Application Suite 9.2 |
| A question and label rules | Labellers need the same definition. | Mandatory | The dataset’s label guide. | Two labels, with one sentence each and an example. | Two people label the same image and agree. | The same crack is labelled two ways. | Write the guide before the bulk of the labelling. | Modules and applications in Maximo Application Suite 9.2 |
| Labelled images you are allowed to use, and a holdout | The model learns the labels. The holdout tests them. | Mandatory | The dataset, with a holdout set aside. | A training folder and a smaller holdout folder, same labels. | Train without the holdout. Score one holdout image. | The only correct result is on an image the model trained on. | Score the holdout. | Modules and applications in Maximo Application Suite 9.2 |
| Storage and a path for an edge device | Images and devices are infrastructure, not an AppPoint formula. | Conditional | Image storage. The network path to an edge device if you use one. | Storage sized for the training set. One device that can pull a model. | Upload the training set. If a device is in scope, it receives the model. | Training stops because the volume is full, or the device never contacts the server. | Extend storage or open the device path. Keep that separate from the 45-point check. | Modules and applications in Maximo Application Suite 9.2 |
A. Ready to configure
- Premium and the 45-point install are both on the entitlement and the dashboard.
- Visual Inspection opens.
- The question and two labels are written.
- You have images you are allowed to label.
B. Ready for the first end-to-end journey
- The dataset trains on the labelled set.
- A holdout image returns a label.
- The label matches the guide or the miss is visible.
- No work order is created unless you add that step on purpose.
C. Ready for production operation
- The label guide has an owner.
- Each deployed camera is counted against the edge install.
- Image retention and access are agreed.
- A wrong label is reviewed as data before the model is blamed, and a missing install line is reviewed as entitlement.
Sources for this readiness check
Licence statements come from the IBM topic named on the row. Installation of a component is a separate check from entitlement.
- Licensing in Maximo Application Suite 9.2 — MAS 9.2 licensing guidance, verified 2026-10-18
- User access and entitlements in MAS 9.1 and later — MAS 9.1 and later, verified 2026-10-18
- Accessing the AppPoints Usage Dashboard — MAS 9.2, verified 2026-10-18
- Modules and applications in Maximo Application Suite 9.2 — MAS 9.2, verified 2026-10-18
Before you start
MAS learning path. Match the component documentation to your installed release and maintenance build before configuring it.
- Visual Inspection available with suitable training resources.
- A permitted synthetic or training image collection with clear labeling rules.
IBM sources + Assetize exercise
01 · Basics — Understand the building blocks
Visual Inspection provides a workflow for creating and training models to inspect assets and products. Training a model and deploying it for use are separate steps.
A useful dataset includes realistic variation in lighting, angles and defects. Near-identical images across training and evaluation can inflate apparent performance.
Topics in this series
- Inspection objective
- Images and labels
- Model training
- Evaluation
- Deployment and inference
02 · Configuration — Work through the setup
Use a training environment and match the actions to the IBM guide for your installed build.
- Define one inspection question and label categories consistently.
- Create an image dataset and review annotations before training.
- Hold out an independent evaluation set, separating near-duplicates from the training set.
- Train using a supported model workflow and inspect mistakes on the held-out images.
- Deploy a selected model and send a new image for inference; retain model identity and the observed response.
03 · Practical example
Fictional teaching example
A teaching dataset distinguishes visibly damaged from intact housings. A model that recognizes the background instead of the defect can appear accurate until the camera location changes.
04 · Practice and verify
Compare inference on a normal image, a defect image and an unfamiliar lighting condition.
Expected result
The deployed model identity is known, and incorrect or uncertain detections are visible in the evaluation record.
Common mistakes
- Reusing training images as proof of generalization.
- Equating a confidence score with guaranteed correctness.
05 · Advanced concepts
Explore confidence thresholds, human review and model replacement. Treat Edge deployment as a separate device/runtime validation when applicable.
IBM references
Product explanations link to IBM sources. Scenarios and exercises are original Assetize teaching material. Reviewed 2026-10-01.
Maximo ecosystem overview · Manage learning paths · Practice data in Workbench