3.0 Configure watsonx.governance

IBM Certified watsonx Governance Lifecycle Advisor v1 - Associate · 60 questions

• When provisioning watsonx.governance from the IBM Cloud Catalog, which three core services must also be provisioned to support it?
• An admin needs to activate full OpenPages integration in watsonx.governance, enabling both AI Factsheets and OpenScale monitoring through a single configuration action. Where is this toggle found?
• A data science team wants to designate their Watson OpenScale deployment space for testing models before go-live. Which environment type should they configure for this deployment space?
• Where in watsonx.governance does an administrator navigate to create a new model inventory for tracking AI assets?
• After deploying a model to a Watson Machine Learning deployment space, what must a governance administrator configure in Watson OpenScale to begin monitoring that specific model?
• Before Watson OpenScale can monitor models hosted on an external ML platform such as Amazon SageMaker, what initial step must be completed inside Watson OpenScale?
• When configuring Watson OpenScale subscriptions, which component stores the model payload logging data (prediction inputs and outputs) used for monitoring calculations?
• A company wants to use their existing corporate directory to manage user access to watsonx.governance, rather than maintaining a separate local user list. Which configuration approach should they select?
• What is the correct sequence to link a model already deployed in a Watson Machine Learning space to Watson OpenScale for active monitoring?
• An OpenPages administrator needs to automate a multi-step AI model approval process — routing for review, collecting sign-offs, and sending notifications — without writing any code. Which OpenPages feature is designed for this purpose?
• In OpenPages, a governance team wants to define a structured hierarchy of AI risk categories such as bias risk, explainability risk, and operational risk for use across all AI use cases. Where is this classification structure configured?
• After creating an AI use case in a watsonx.governance model inventory, how are AI Factsheets associated with that use case to provide lifecycle traceability?
• An organization wants to rapidly implement a predefined AI governance framework in OpenPages that includes pre-built controls, regulatory mappings, and risk templates aligned to industry standards. Which feature enables this?
• A model governance team wants Watson OpenScale to generate an alert whenever a monitored model's fairness metric falls below 80%. Where must this threshold be configured?
• A compliance officer in OpenPages needs to gather structured risk assessment data from AI model owners about a new high-risk use case. Which OpenPages feature is designed to collect this information from multiple stakeholders?
• A governance team wants to capture organization-specific information in AI Factsheets, such as the training data source location and model owner's department. How is this accomplished in watsonx.governance?
• A governance team wants AI Factsheets to automatically record a lifecycle event when a model is promoted from the development stage to production. Which configuration enables this behavior?
• A machine learning model trained and deployed on Amazon SageMaker must be governed and monitored through watsonx.governance. What configuration step enables Watson OpenScale to actively monitor this external model?
• A data scientist trains a model in a watsonx.ai project that has been linked to a governed AI use case in watsonx.governance. What happens automatically as a result of this integration?
• When configuring Watson OpenScale to monitor models, an administrator must connect it to the deployment environment where the models are hosted. How is this connection established?
• Which option correctly describes the organizational hierarchy of the watsonx.governance model inventory structure?
• To enable Watson OpenScale's quality monitors to measure a model's real-world accuracy over time using labeled production outcomes, what must be configured within the model's subscription?
• In OpenPages, an administrator wants to ensure that only members of the Risk Management group can approve AI use case risk assessments while all other users have read-only access. How is this access control enforced?
• A governance officer needs a single consolidated view in OpenPages showing all AI models currently under review, their risk scores, and current workflow status. Which OpenPages capability provides this view?
• A DevOps team wants to programmatically create AI use cases and register governed model assets in watsonx.governance as part of their automated CI/CD pipeline. Which approach is correct?
• An administrator setting up Watson OpenScale for the first time can choose between automated and manual configuration modes. What is the key functional difference between these two paths?
• A pilot deployment's governance platform needs its policies, questionnaires, and workflows to actually reflect the organisation's obligations. What should be configured?
• The governance platform currently relies on people to gather evidence about a model by hand. Which choice best lets it collect that evidence automatically instead?
• A team is preparing to roll out watsonx.governance across the organisation. What should they set up first so AI use cases do not slip through ungoverned?
• A senior engineer reviewing a governance plan notices every project member has been made an owner of the project's models. What principle should drive the correction?
• Rolling out governance across many business units, a customer proposes leaving every questionnaire at its default text to save configuration time. What is the risk in this approach?
• Turning a proof of concept into a governed service requires deciding which environments to connect to the governance platform. What is the correct scope?
• An administrator notices that several AI use cases from a pilot never made it into any governance record. What is the most likely root cause?
• A governance workflow needs to follow the principle of least privilege. Which choice best matches that principle?
• Ahead of a production rollout, someone proposes disabling several required questionnaire questions for a high-risk model to speed up sign-off. What should the team do instead to avoid the model slipping through governance with incomplete evidence?
• Each week, metrics are copied from the model platform into the governance system by hand. What risk does that manual process create?
• A customer wants a governance approach that scales as their AI portfolio grows from a handful of models to hundreds. What should a consultant recommend they build first?
• A project lead turning a proof of concept into a governed deployment wants to prevent the same person from building a model and also approving it for production. What principle addresses this?
• A pilot deployment's policy was configured once during the initial demo and never revisited. What is the concern with treating that setup as final?
• Treating governance as entirely separate from the model development and deployment tooling would be a mistake. What is the reason?
• Which finding would indicate that a team's governance setup is working as intended during a production rollout?
• A senior engineer reviews a plan that has one shared login used by everyone on the governance team. What is the main risk this creates?
• A customer needs guidance on how to structure workflow requirements for model approvals. Which structure best matches how governance obligations actually work?
• A project lead turning a proof of concept into a governed service is told that connecting governance tooling to the development environment is unnecessary once the model is in production. Why is that advice wrong?
• During a pilot, an administrator wants to confirm that governance workflows are correctly wired up before scaling to more use cases. What should they check?
• A new team member needs a governance role assigned. What should determine which one they receive?
• Ahead of a production rollout, the questionnaire assigned to a model asks about data retention practices that have nothing to do with the organisation's actual policy. What does this indicate?
• A senior engineer reviewing the current plan asks what governance tooling should do once it is connected to development and deployment environments. What is its role?
• A customer wants to know why governance inventories should be configured before onboarding any AI use cases, rather than after. What is the best justification?
• A project lead notices that a model builder still has permission to delete approval records after the model has gone live. Which principle has been violated?
• An administrator troubleshooting policy configuration finds that two similar high-risk models have been assigned different sets of required questions with no documented reason. What should be reviewed first?
• Which setup best reflects the integration an architect should design between governance tooling and a team's model environments?
• A team preparing a production rollout is deciding how to onboard a new generative AI use case into governance. Which approach fits the intended design?
• Which combination best summarises how governance roles should be structured across a project?
• Which outcome shows that policy, questionnaire, and workflow configuration has been done correctly?
• What is the primary purpose of connecting governance tooling with model development and deployment environments?
• Which statement best summarises why configuring inventories and workflows is the starting point for watsonx.governance adoption?
• A governance administrator account has full rights across every project. What control best reduces the risk this account represents?
• A team preparing a production rollout is deciding how to configure governance for their model. Which statement best captures the correct approach?
• Summing up, why should governance tooling be connected to both the development and deployment environments rather than treated as a standalone system fed by occasional manual updates?