Deploying Maximo AI Service

You can use Maximo® AI Service SaaS or on-premises. To deploy Maximo AI Service on-premises, you initiate the deployment by using the CLI or Ansible® collection, connect Maximo AI Service to Maximo Manage by using system properties, and then conditionally import the Maximo AI Service certificate.

Before you begin

Before you begin, review the entire process for enabling AI features. For more information, see Maximo AI Service and AI features in Maximo Manage.

You must also set up prerequisites. The following list describes the prerequisites:
  • Install or set up an account for watsonx™.ai. You can use watsonx.ai on-premises or SaaS. You can use an existing instance or install or sign up for a new instance.

  • At a minimum, for a single user and tenant, you must have three primary nodes that have eight CPUs with 32 GB memory each and six secondary nodes that have four CPUs with 32 GB memory each. Additional users, tenants, and workloads require more resources.

  • Red Hat® OpenShift® 4.16 or later

  • Suite License Service. You can use an existing instance or install a new instance as part of the deployment for Maximo AI Service.

  • IBM® Data Reporter Operator . You can use an existing instance or install a new instance as part of the deployment for Maximo AI Service.

  • One of the following object storage providers:

    • MinIO. You can use an existing instance or install a new instance as part of the deployment for Maximo AI Service. Install MinIO in the same cluster as Maximo AI Service.

    • Amazon Web Services S3. The buckets that are used for Maximo AI Service must have unique names, and all typical dependencies of Amazon Web Services S3 must be deployed.

  • IBM Db2®. You can use an existing instance or install a new instance as part of the deployment for Maximo AI Service.

About this task

As part of the deploying, you use system properties to establish a connection from Maximo Manage to the Maximo AI Service. This connection enables Maximo Manage to communicate with the services or runtimes that are hosted within the Maximo AI Service cluster.

The following steps describe how to deploy Maximo AI Service on-premises. For more information about Maximo AI Service SaaS, see Enabling AI features by using Maximo AI Service SaaS.

Procedure

  1. Install Maximo AI Service by using the Maximo Application Suite CLI or the Ansible collection.

    For more information about deploying by using Ansible, see Installing Maximo AI Service with the Ansible collection.

    For more information about deploying by using the CLI in interactive mode, see Installing Maximo AI Service from the CLI in interactive mode.

    For more information about deploying by using the CLI in noninteractive mode, see Installing Maximo AI Service from the CLI in noninteractive mode.

  2. Retrieve the required values for connecting Maximo Manage to Maximo AI Service.
    You must retrieve the Maximo AI Service API key, URL, and tenant ID. All of these values are in Red Hat OpenShift web console. You must have these values to complete step 3.
    1. Retrieve the API key.
      1. In Red Hat OpenShift web console, click Home > Projects and search for the aiservice-instance_id project name.
      2. Open the aiservice-instance_id project and in the Inventory section, click number Secrets.
      3. Click the aiservice-user----apikey-secret secret and in the Data section, click Reveal values. Save this value on your local machine.
    2. Retrieve the URL.
      1. In Red Hat OpenShift web console, click Networking > Routes and search for the aibroker route.
      2. Open the route.
      3. Copy the value that is in the Location field.
      4. On your local machine, save that URL and append the following path to the end of the URL: /ibm/aibroker/service/rest/api/v1
    3. Retrieve the tenant ID.
      1. In Red Hat OpenShift web console, click Administration > CustomResourceDefinition and search for the AIServiceTenant definition name.
      2. Open the definition.
      3. On the Instances tab, copy the value in the Name column that corresponds to the system that you are setting up. For example, if you are deploying Maximo AI Service in your production environment, copy the Name column value for your production tenant.
  3. Configure the Maximo Manage system properties.
    1. In Maximo Manage, open the System Properties application.
    2. Search for and then select and add global values for the following properties. Use the values that you retrieved in step 2.
      • mxe.int.aibrokerapikey. The value is the Maximo AI Service API key.
      • mxe.int.aibrokerapiurl. The value is the Maximo AI Service URL.
      • mxe.int.aibrokertenantid. The value is the Maximo AI Service tenant ID.
    3. After you edit each property, in the Common Actions menu, click Save Property.
    4. After you edit all properties, in the Common Actions menu, click Live Refresh.
  4. Create and import the ca.cert file for Maximo AI Service.
    Some systems do not require importing the certificate. You can skip this step and first verify that Maximo AI Service is running. If Maximo AI Service is not running, you can set up the certificate. If Maximo AI Service is running, then the certificate is not needed.
    1. In Red Hat OpenShift web console, Home > Projects and search for the mas-instance_id-broker project.
    2. Open the project.
    3. From the side navigation menu, click Workloads > Secrets.
    4. Select the instance_id-public-aibroker-tls secret.
    5. Copy the content that is in the ca.crt field. The value for the ca.crt field is the content for the ca.cert file.
    6. On your local machine, paste the content into an empty .txt file.
    7. Save the file as ca.cert.
    8. Import the certificate. For more information, see Adding trusted certificates.
  5. Verify that Maximo AI Service is running and connected to Maximo Manage.
    To verify that Maximo AI Service is running and connected, you can run the following command:
    curl -X 'GET' \ 'https://{hostname}/ibm/aibroker/service/rest/api/v1/health' \ -H 'accept: */*'x
    If Maximo AI Service is running, the following output is returned: {"max_number_of_tenant":"<number>","kmodel":"running","healthy":true,"version":"<version>","status":{"KMODELS":{"healthy":true},"DB2":{"healthy":true}}}
    Alternately, you can check the status of Maximo AI Service in Maximo Manage in the AI configuration application. In the AI configuration application, click AI Service Health. If the status is running, Maximo AI Service is ready to use.

What to do next

If another status or an error is displayed, you can access more details in the Red Hat OpenShift logs. For more information about troubleshooting, see Troubleshooting Maximo AI Service and AI features.

If Maximo AI Service is available and running, you can start setting up your AI configurations. Each configuration represents a feature that you want to enable, for example, problem code recommendations or the AI assistant.

The following table contains the available AI features, associated model template name, the required Granite™ model, and links to steps to set up the AI configuration.
Table 1. AI features and enablement
Feature Model template Model Steps

Problem code recommendations for work orders

pcc

 Granite 3.0 8B Instruct Enabling recommended problem codes for Work orders

Field value recommendations

mcc

 Granite 3.2 8B Instruct Enabling field value recommendations

AI assistant

nl2oslc

 Granite 3.2 8B Instruct Enabling the assistant

Locating similar work orders

similarity

 Granite 3.0 8B Instruct Enabling locating of similar work orders

AI recommendations for asset boundary and failure list in Reliability Strategies

fmea

 Granite 3.2 8B Instruct Enabling AI recommendations in Reliability Strategies