name: mbr-engine version: 2.0.0 description: Monthly Business Review automation - extract org health metrics, detect patterns, generate PowerPoint presentations author: jac007x created: 2026-03-18 status: in_development

Compliance

compliance: review_date: 2026-03-18 reviewer: jac007x policies: - AI-01-02 - DG-01-ST-02 - Ethical AI Principles status: compliant risk_level: medium pii_handling: true next_review: 2026-06-18

Model Recommendation

model_recommendation: sonnet model_rationale: > Pattern detection, metric analysis, and slide narrative generation require moderate complexity. Sonnet provides the analytical depth needed while being more efficient than Opus. Haiku insufficient for feature detection.

tags:

• hr-analytics
• reporting
• automation
• powerpoint
• org-health
• metrics

Dependencies

requires: python: ">=3.10" packages: - pandas>=2.0 - openpyxl>=3.1 - python-pptx>=0.6.21

Inputs

inputs:

• name: roster_file type: file format: xlsx description: HR roster export with employee data required: true

• name: finance_file type: file format: xlsx description: Finance HC vs AOP data required: false

• name: mada_file type: file format: xlsx description: Recognition budget data (MADA) required: false

• name: month type: string format: "YYYY-MM" description: Reporting month default: current_month

Outputs

outputs:

• name: mbr_slides type: file format: pptx description: Generated PowerPoint presentation

• name: metrics_json type: file format: json description: Extracted metrics in JSON format

• name: feature_topics type: list description: Detected patterns ranked by importance

Components

components:

• name: delta description: Data extraction layer modules:

  • org_health_extractor
  • headcount_extractor
  • turnover_extractor
  • recognition_extractor

• name: feature_engine description: Pattern detection and scoring modules:

  • pattern_detector
  • topic_scorer
  • topic_memory

• name: pptx_builder description: PowerPoint generation modules:

  • slide_builder
  • table_formatter
  • chart_creator

• name: threshold_governance description: Metric threshold management modules:

  • threshold_definitions
  • threshold_review_engine
  • persistence_layer

• name: portability description: Multi-org support modules:

  • org_registry
  • schema_registry
  • metric_registry
  • threshold_discovery
  • org_onboarding

Configuration

config: orgs: description: Organizations to include in MBR type: list default: [Tech, Services, AIPD]

feature_topic_count: description: Max feature topics to recommend type: integer default: 5

persist_metrics: description: Save metrics for MoM tracking type: boolean default: true

threshold_review_months: description: Months to run threshold reviews type: list default: [2, 5, 8, 11] # Feb, May, Aug, Nov

Workflows

workflows:

• name: full_mbr description: Complete MBR generation workflow steps:

  • extract_org_health
  • extract_headcount
  • extract_recognition
  • detect_features
  • generate_slides

• name: quarterly_review description: Quarterly threshold review steps:

  • load_historical_data
  • analyze_achievability
  • generate_recommendations
  • create_review_report

• name: onboard_org description: Add a new organization steps:

  • data_discovery
  • context_gathering
  • sample_extraction
  • threshold_calibration
  • validation_review

Metrics produced

metrics:

• id: ic_manager_ratio name: IC:Manager Ratio category: org_structure

• id: avg_span_of_control name: Avg Span of Control category: org_structure

• id: max_layers name: Max Org Layers category: org_structure

• id: mgrs_small_span_pct name: Managers with Small Spans category: org_structure

• id: turnover_total name: Total Turnover category: people

• id: turnover_voluntary name: Voluntary Turnover category: people

• id: recognition_utilization_pct name: Recognition Utilization category: recognition

Portability

portability: supports_custom_orgs: true supports_custom_metrics: true supports_custom_schemas: true onboarding_required: true schema_contribution: true