Domain I: Support Responsible and Trustworthy AI Efforts - 15%
|
|
Task 1 |
- Oversee privacy and security plan:
-
Establish data governance protocols for personally identifiable information (PII)
-
Implement encryption and access controls for AI training data
-
Conduct privacy impact assessments for AI model deployment
-
Ensure compliance with GDPR, CCPA, and other data protection regulations
-
Design secure data handling procedures throughout the AI lifecycle
|
|
Task 2 |
- Manage AI/ML transparency (e.g., data selection, algorithm selection):
-
Document model selection criteria and decision rationale
-
Create transparent reporting on data sources and preprocessing steps
-
Establish explainability requirements for stakeholder communication
-
Maintain audit trails for algorithmic decision-making processes
-
Implement model interpretability tools and techniques
|
|
Task 3 |
- Conduct bias checks (e.g., model, data, algorithm):
-
Analyze training data for demographic and representation imbalances
-
Perform fairness testing across different population groups
-
Implement bias detection metrics and monitoring systems
-
Review model outputs for discriminatory patterns
-
Apply bias mitigation techniques during model development
|
|
Task 4 |
- Monitor regulatory and policy compliance:
-
Track evolving AI regulations and industry standards
-
Ensure adherence to sector-specific compliance requirements
-
Coordinate with legal and compliance teams on AI governance
-
Implement compliance monitoring and reporting mechanisms
-
Maintain documentation for regulatory audits and reviews
|
|
Task 5 |
- Manage accountability documentation and audit trail:
-
Create comprehensive records of AI model development decisions
-
Establish version control for models, data, and training processes
-
Document stakeholder approvals and go/no-go decision points
-
Maintain chain of custody records for training and test data
-
Prepare accountability reports for executive and regulatory review
|
Domain II: Identify Business Needs and Solutions - 26%
|
|
Task 1 |
- Identify problem to be solved (e.g., needs, persona)
-
Conduct stakeholder interviews to understand business pain points
-
Analyze existing processes to identify automation opportunities
-
Define target user personas and use cases for AI solutions
-
Map business problems to appropriate AI patterns and approaches
-
Validate problem statements with subject matter experts
|
|
Task 2 |
- Evaluate initial AI feasibility
-
Assess technical viability of proposed AI solutions
-
Analyze data availability and quality for model training
-
Evaluate computational resource requirements and constraints
-
Review organizational readiness for AI implementation
-
Compare AI approaches against traditional solution alternatives
|
|
Task 3 |
- Conduct risk assessment(s) (e.g., security, safety, ethics)
-
Identify potential failure modes and safety implications
-
Assess cybersecurity vulnerabilities in AI systems
-
Evaluate ethical implications of AI decision-making
-
Analyze reputational and business continuity risks
-
Develop risk mitigation strategies and contingency plans
|
|
Task 4 |
- Develop AI project scope statement
-
Define project boundaries and deliverables for AI initiatives
-
Establish success criteria and performance metrics
-
Identify in-scope and out-of-scope functionality
-
Document assumptions and constraints for AI implementation
-
Align scope with business objectives and resource availability
|
|
Task 5 |
- Determine ROI
-
Calculate expected benefits from AI solution implementation
-
Estimate total cost of ownership including infrastructure and maintenance
-
Develop business case with financial justification
-
Establish metrics for measuring return on investment
-
Create cost-benefit analysis for stakeholder decision-making
|
|
Task 6 |
- Manage adoption/integration risks
-
Assess organizational change management requirements
-
Identify potential user resistance and adoption barriers
-
Plan integration with existing systems and workflows
-
Develop training and communication strategies for end users
-
Monitor adoption metrics and address implementation challenges
|
|
Task 7 |
- Draft AI solution
-
Create high-level architecture for AI system design
-
Define data flow and processing requirements
-
Specify AI model types and algorithmic approaches
-
Document integration points with existing systems
-
Outline deployment and operational considerations
|
|
Task 8 |
- Define success criteria (e.g., KPIs, metrics)
-
Establish measurable performance indicators for AI models
-
Define business impact metrics and success thresholds
-
Create technical performance benchmarks and targets
-
Develop user satisfaction and adoption measurement criteria
-
Align success metrics with organizational objectives
|
|
Task 9 |
- Support business case creation
-
Gather financial data and projected benefits for business case
-
Collaborate with finance teams on cost estimates and projections
-
Develop compelling narratives for executive presentations
-
Provide technical expertise for business case validation
-
Review and refine business case documentation
|
|
Task 10 |
- Identify project resources (e.g., people, hardware, contractors)
-
Assess skill requirements for AI project team composition
-
Evaluate hardware and infrastructure needs for development and deployment
-
Identify gaps requiring external contractors or consultants
-
Plan resource allocation and timeline for project phases
-
Coordinate with procurement for specialized AI tools and platforms
|
Domain III: Identify Data Needs - 26%
|
|
Task 1 |
- Define required data
-
Specify data types and formats needed for AI model training
-
Determine data volume requirements and sampling strategies
-
Identify temporal and granularity requirements for data collection
-
Define data quality standards and acceptance criteria
-
Map data requirements to business objectives and use cases
|
|
Task 2 |
- Identify data SMEs
-
Locate domain experts with knowledge of relevant data sources
-
Engage business users who understand data context and meaning
-
Connect with data stewards and data governance teams
-
Identify technical experts familiar with data systems and structures
-
Establish communication channels with identified subject matter experts
|
|
Task 3 |
- Identify data sources and locations
-
Map internal databases and data warehouses containing relevant information
-
Explore external data sources and third-party data providers
-
Assess cloud storage and distributed data repositories
-
Inventory legacy systems and historical data archives
-
Document data ownership and access permissions
|
|
Task 4 |
- Coordinate AI workspace and infrastructure
-
Provision computing resources for data processing and model training
-
Establish secure development environments for AI teams
-
Configure data storage and backup systems for project needs
-
Set up collaboration tools and version control systems
-
Ensure compliance with security and governance requirements
|
|
Task 5 |
- Gather required data
-
Execute data extraction from identified sources and systems
-
Coordinate data transfers and migrations to AI development environments
-
Implement data collection processes for ongoing data feeds
-
Validate data completeness and accuracy during collection
-
Establish data refresh and update procedures
|
|
Task 6 |
- Check data privacy, compliance, and access
-
Verify data usage rights and licensing agreements
-
Ensure compliance with data protection regulations and policies
-
Implement access controls and user permissions for data resources
-
Conduct privacy impact assessments for data usage
-
Document data lineage and usage for audit purposes
|
|
Task 7 |
- Oversee data evaluation
-
Assess data quality dimensions including accuracy, completeness, and consistency
-
Analyze data distributions and identify potential biases or gaps
-
Evaluate data freshness and relevance for AI model training
-
Review data schema and structure for modeling compatibility
-
Conduct exploratory data analysis to understand data characteristics
|
|
Task 8 |
- Determine if data meets solution needs
-
Compare available data against defined requirements and specifications
-
Assess data sufficiency for training robust AI models
-
Identify data gaps and develop strategies for addressing deficiencies
-
Validate data representativeness for target use cases
-
Make go/no-go decisions based on data readiness assessment
|
|
Task 9 |
- Convey data understanding to leadership
-
Prepare executive summaries of data assessment findings
-
Create visualizations and reports to communicate data insights
-
Present data readiness status and recommendations to stakeholders
-
Translate technical data concepts into business-relevant language
-
Provide regular updates on data preparation progress and challenges
|
Domain IV: Manage AI Model Development and Evaluation - 16%
|
|
Task 1 |
- Oversee AI/ML model technique(s) (e.g., algorithm, selection)
-
Research and evaluate appropriate algorithms for specific use cases
-
Guide selection between supervised, unsupervised, and reinforcement learning approaches
-
Assess trade-offs between model complexity, performance, and interpretability
-
Coordinate with data scientists on model architecture decisions
-
Review algorithm selection criteria and decision documentation
|
|
Task 2 |
- Oversee AI/ML model QA/QC (e.g., configuration management, model performance)
-
Establish model testing protocols and quality assurance procedures
-
Implement configuration management for model versions and parameters
-
Monitor model performance metrics during development and testing
-
Coordinate peer reviews and technical validation of model designs
-
Ensure adherence to coding standards and best practices
|
|
Task 3 |
- Manage AI/ML model training
-
Plan training schedules and resource allocation for model development
-
Monitor training progress and computational resource utilization
-
Coordinate hyperparameter tuning and optimization activities
-
Oversee cross-validation and model selection processes
-
Manage training data versioning and experiment tracking
|
|
Task 4 |
- Manage data transformation to conduct data preparation
-
Oversee data cleaning and preprocessing workflows
-
Coordinate feature engineering and selection activities
-
Manage data normalization and standardization processes
-
Supervise data augmentation and synthetic data generation
-
Ensure data transformation reproducibility and documentation
|
|
Task 5 |
- Verify data quality for go/no-go decision to conduct data preparation
-
Conduct final data quality assessments before model training
-
Validate data preprocessing and transformation results
-
Assess data representativeness and potential bias issues
-
Make decisions on data readiness for model development
-
Document data quality findings and recommendations
|
|
Task 6 |
- Verify model ready for operationalization go/no-go decision
-
Evaluate model performance against established success criteria
-
Assess model robustness and generalization capabilities
-
Review deployment readiness including infrastructure requirements
-
Validate model documentation and operational procedures
-
Make final approval decisions for model deployment
|
Domain V: Operationalize AI Solution - 17%
|
|
Task 1 |
- Manage creation of AI solution deployment plan
-
Develop comprehensive deployment strategy and timeline
-
Plan infrastructure requirements and resource allocation
-
Coordinate with IT teams on system integration and deployment
-
Establish rollback procedures and contingency plans
-
Create deployment checklists and validation criteria
|
|
Task 2 |
- Manage AI solution deployment
-
Coordinate deployment activities across technical teams
-
Monitor deployment progress and resolve implementation issues
-
Validate system functionality and performance in production environment
-
Manage user access provisioning and security configurations
-
Conduct post-deployment verification and testing
|
|
Task 3 |
- Oversee model governance
-
Establish model lifecycle management procedures
-
Implement model versioning and change control processes
-
Monitor model performance and drift detection
-
Coordinate model updates and retraining schedules
-
Ensure compliance with governance policies and standards
|
|
Task 4 |
- Oversee AI solution metrics (e.g., KPI, model performance)
-
Implement monitoring dashboards for business and technical metrics
-
Track key performance indicators and success measures
-
Analyze model performance trends and degradation patterns
-
Generate regular performance reports for stakeholders
-
Establish alerting systems for performance threshold breaches
|
|
Task 5 |
- Prepare final report/lessons learned
-
Document project outcomes and achievement of objectives
-
Capture lessons learned and best practices for future projects
-
Analyze what worked well and areas for improvement
-
Create knowledge transfer documentation for operational teams
-
Present final project results to stakeholders and leadership
|
|
Task 6 |
- Manage AI solution transition plan
-
Plan transition from project team to operational support
-
Coordinate knowledge transfer to production support teams
-
Establish ongoing maintenance and support procedures
-
Define roles and responsibilities for operational phase
-
Create handover documentation and training materials
|
|
Task 7 |
- Oversee AI solution contingency plan
-
Develop incident response procedures for AI system failures
-
Plan backup and disaster recovery strategies
-
Establish escalation procedures for critical issues
-
Create business continuity plans for AI service disruptions
-
Test and validate contingency procedures regularly
|
To achieve the professional designation of PMI Certified Professional in Managing AI from the PMI, candidates must clear the PMI-CPMAI Exam with the minimum cut-off score. For those who wish to pass the PMI Managing AI Professional certification exam with good percentage, please take a look at the following reference document detailing what should be included in PMI Managing AI Professional Exam preparation.
