SuperAPI AI Directory

cwm-tasks-generation

General & other

View repository

Updated 10-16 05:50

About this skill

CWM Task Generation

Overview

This skill covers generating structured task breakdowns from requirements, user stories, or high-level project descriptions in ServiceNow CWM:

  • Decomposing requirements into hierarchical work breakdown structures (WBS)
  • Creating project tasks with effort estimates, durations, and skill requirements
  • Defining task dependencies (finish-to-start, start-to-start, finish-to-finish)
  • Assigning tasks to team members based on skills and availability
  • Setting milestones and phase gates at appropriate project checkpoints
  • Generating parallel workstreams where tasks can execute concurrently

When to use: When starting a new project phase, converting approved requirements into executable tasks, breaking down epics or user stories, or restructuring a project plan for replanning.

Prerequisites

  • Roles: project_manager, program_manager, or admin
  • Plugins: com.snc.project_management (Project Portfolio Management) or CWM workspace
  • Access: Write access to pm_project_task, planned_task tables
  • Data: An existing project record and defined requirements or user stories
  • Related Skills: spm/cwm-doc-generation for status reporting, spm/acceptance-criteria for defining done criteria

Procedure

Step 1: Analyze the Source Requirements

Review the input requirements or user stories to identify work components.

MCP Approach:

Tool: SN-Get-Record
Parameters:
  table_name: pm_project
  sys_id: [PROJECT_SYS_ID]
  fields: sys_id,number,short_description,description,planned_start_date,planned_end_date,project_manager,methodology

If requirements are stored as stories or demand items:

Tool: SN-Query-Table
Parameters:
  table_name: rm_story
  query: product=[PROJECT_SYS_ID]^state!=cancelled
  fields: sys_id,number,short_description,description,acceptance_criteria,story_points,priority
  limit: 50

Decomposition Framework:

Requirement Type Decomposition Approach
Epic / Feature Break into user stories, then tasks
User Story Break into implementation tasks (design, build, test, deploy)
Technical Requirement Break into analysis, implementation, validation tasks
Business Process Break into process steps, each becoming a task
Integration Break into design, develop, test, certify tasks

Step 2: Define Work Breakdown Structure

Create the hierarchical task structure with parent-child relationships.

PROJECT: [Project Name]
|
+-- Phase 1: [Phase Name]
|   +-- WP 1.1: [Work Package]
|   |   +-- Task 1.1.1: [Task]
|   |   +-- Task 1.1.2: [Task]
|   +-- WP 1.2: [Work Package]
|   |   +-- Task 1.2.1: [Task]
|   +-- Milestone: [Phase 1 Complete]
|
+-- Phase 2: [Phase Name]
|   +-- WP 2.1: [Work Package]
|   ...

Step 3: Create Phase Tasks

Build the top-level phase containers.

MCP Approach:

Tool: SN-Create-Record
Parameters:
  table_name: pm_project_task
  data:
    project: [PROJECT_SYS_ID]
    short_description: "Phase 1: Requirements and Design"
    description: "Gather detailed requirements, create solution design, and obtain stakeholder approval"
    state: pending
    priority: 2
    planned_start_date: "2026-04-01"
    planned_end_date: "2026-04-30"
    top_task: true
    wbs: "1"

REST Approach:

POST /api/now/table/pm_project_task
Body: {
  "project": "[PROJECT_SYS_ID]",
  "short_description": "Phase 1: Requirements and Design",
  "description": "Gather detailed requirements, create solution design, and obtain stakeholder approval",
  "state": "pending",
  "priority": "2",
  "planned_start_date": "2026-04-01",
  "planned_end_date": "2026-04-30",
  "top_task": "true",
  "wbs": "1"
}

Step 4: Create Work Package and Leaf Tasks

Build detailed tasks under each phase.

MCP Approach:

Tool: SN-Create-Record
Parameters:
  table_name: pm_project_task
  data:
    project: [PROJECT_SYS_ID]
    parent: [PHASE_TASK_SYS_ID]
    short_description: "Conduct stakeholder interviews"
    description: "Schedule and conduct interviews with 8 key stakeholders to gather requirements for the new system"
    state: pending
    priority: 2
    planned_start_date: "2026-04-01"
    planned_end_date: "2026-04-05"
    estimated_work_duration: "3 days"
    assigned_to: [USER_SYS_ID]
    assignment_group: [GROUP_SYS_ID]
    wbs: "1.1"
    skill: "Business Analysis"

REST Approach:

POST /api/now/table/pm_project_task
Body: {
  "project": "[PROJECT_SYS_ID]",
  "parent": "[PHASE_TASK_SYS_ID]",
  "short_description": "Conduct stakeholder interviews",
  "state": "pending",
  "planned_start_date": "2026-04-01",
  "planned_end_date": "2026-04-05",
  "estimated_work_duration": "3 days",
  "assigned_to": "[USER_SYS_ID]",
  "wbs": "1.1"
}

Standard Task Templates per Requirement Type:

User Story Task Est. Effort Skills Needed
Requirements Analysis 4-8 hrs Business Analyst
Solution Design 8-16 hrs Architect/Tech Lead
Implementation/Build 16-40 hrs Developer
Unit Testing 4-8 hrs Developer
Integration Testing 8-16 hrs QA Engineer
User Acceptance Testing 4-8 hrs BA + End Users
Documentation 4-8 hrs Tech Writer/BA
Deployment 2-4 hrs DevOps/Admin

Step 5: Define Task Dependencies

Create predecessor/successor relationships between tasks.

MCP Approach:

Tool: SN-Update-Record
Parameters:
  table_name: pm_project_task
  sys_id: [TASK_SYS_ID]
  data:
    predecessor: [PREDECESSOR_TASK_SYS_ID]
    predecessor_type: "FS"

Dependency Types:

Type Code Meaning Example
Finish-to-Start FS B starts after A finishes Build after Design
Start-to-Start SS B starts when A starts Testing starts with Build
Finish-to-Finish FF B finishes when A finishes Documentation finishes with Build
Start-to-Finish SF B finishes when A starts Rarely used

Common Dependency Patterns:

Requirements Analysis (FS) -> Solution Design (FS) -> Implementation
Implementation (FS) -> Unit Testing (FS) -> Integration Testing
Integration Testing (FS) -> UAT (FS) -> Deployment
Documentation (SS) -> Implementation (parallel start)

Step 6: Set Milestones

Create milestone tasks at key project checkpoints.

MCP Approach:

Tool: SN-Create-Record
Parameters:
  table_name: pm_project_task
  data:
    project: [PROJECT_SYS_ID]
    parent: [PHASE_TASK_SYS_ID]
    short_description: "MILESTONE: Design Review Approved"
    state: pending
    milestone: true
    planned_end_date: "2026-04-15"
    duration: "0"
    predecessor: [DESIGN_TASK_SYS_ID]
    wbs: "1.5"

Recommended Milestones:

Milestone Trigger Gate Criteria
Requirements Approved All requirements reviewed Sign-off from sponsor
Design Complete Architecture and design reviewed Technical review passed
Development Complete All features built Code review passed, unit tests green
Testing Complete All test cycles done No P1/P2 defects open
Go-Live Ready Deployment validated Stakeholder approval obtained

Step 7: Estimate Effort and Duration

Apply estimation techniques to each task.

Estimation Methods:

Method When to Use Approach
Analogous Similar past project exists Compare to historical task durations
Parametric Quantifiable work units Multiply unit count by per-unit effort
Three-Point Uncertainty is high (Optimistic + 4*Likely + Pessimistic) / 6
Expert Judgment Novel or complex work Consult SME for estimate

Apply estimates to tasks:

Tool: SN-Update-Record
Parameters:
  table_name: pm_project_task
  sys_id: [TASK_SYS_ID]
  data:
    estimated_work_duration: "16 hours"
    planned_start_date: "2026-04-07"
    planned_end_date: "2026-04-09"
    work_notes: "Estimate based on three-point: Opt=12h, Likely=16h, Pess=24h. PERT=16.3h"

Step 8: Assign Resources

Match tasks to team members based on skills and availability.

MCP Approach:

Tool: SN-Query-Table
Parameters:
  table_name: pm_resource_plan
  query: project=[PROJECT_SYS_ID]^state=active
  fields: sys_id,resource,role,planned_hours,actual_hours,remaining_hours,start_date,end_date
  limit: 30
Tool: SN-Update-Record
Parameters:
  table_name: pm_project_task
  sys_id: [TASK_SYS_ID]
  data:
    assigned_to: [USER_SYS_ID]
    assignment_group: [GROUP_SYS_ID]

Step 9: Validate the Task Breakdown

Review the generated task structure for completeness and feasibility.

Validation Checklist:

TASK BREAKDOWN VALIDATION:
- [ ] All requirements have corresponding tasks
- [ ] No orphan tasks (every task has a parent or is top-level)
- [ ] Dependencies form a valid DAG (no circular dependencies)
- [ ] Critical path identified and reasonable
- [ ] All tasks have effort estimates
- [ ] All tasks have assigned resources
- [ ] Milestones placed at phase boundaries
- [ ] Resource loading does not exceed capacity
- [ ] Buffer/contingency included (10-20% of total effort)
- [ ] Task descriptions are clear and actionable

MCP Approach to verify:

Tool: SN-Query-Table
Parameters:
  table_name: pm_project_task
  query: project=[PROJECT_SYS_ID]^assigned_to=NULL^milestone=false
  fields: number,short_description,wbs,state
  limit: 100

Step 10: Generate Task Summary Report

Produce a summary of the generated task breakdown.

=== TASK BREAKDOWN SUMMARY ===
Project: [number] - [name]
Generated: [date]

STRUCTURE:
Phases: [count]
Work Packages: [count]
Tasks: [count]
Milestones: [count]

EFFORT SUMMARY:
Total Estimated Effort: [hours] hours
Total Duration: [days] business days
Critical Path Duration: [days] business days

RESOURCE ALLOCATION:
| Resource | Role | Allocated Hours | Utilization |
|----------|------|----------------|-------------|
| [name] | [role] | [hours] | [%] |

DEPENDENCY MAP:
[Task A] -> [Task B] -> [Task C] (Critical Path)
[Task D] -> [Task E] (Parallel Stream)

MILESTONES:
| Milestone | Date | Dependencies |
|-----------|------|-------------|
| [name] | [date] | [predecessor tasks] |

Tool Usage

Tool Purpose When to Use
SN-Query-Table Find existing tasks, resources, requirements Discovery and validation
SN-Create-Record Create project tasks and milestones Building the task breakdown
SN-Update-Record Set dependencies, assign resources, update estimates Task configuration
SN-Get-Record Get project details and individual task records Context gathering
SN-Get-Table-Schema Discover pm_project_task fields Initial setup

Best Practices

  1. Decompose to 8-40 hour tasks -- tasks smaller than 8 hours create overhead, larger than 40 hours lack visibility
  2. Use verb-noun naming -- "Design database schema" not "Database schema"
  3. Include all work types -- do not forget testing, documentation, deployment, and review tasks
  4. Add buffer tasks -- include 10-20% contingency for unknowns
  5. Define "done" criteria -- each task should have clear completion criteria
  6. Minimize dependencies -- the fewer dependencies, the more scheduling flexibility
  7. Identify critical path -- know which tasks directly impact the project end date
  8. Balance resource loading -- avoid over-allocating any single team member
  9. Review with the team -- validate estimates and assignments with the people doing the work
  10. Version your breakdown -- save baseline before making changes for comparison

Troubleshooting

Issue Cause Resolution
Tasks not rolling up to parent Parent reference not set Verify parent field on child tasks points to correct parent sys_id
Duration calculation incorrect Work calendar not configured Check project work schedule; verify business days vs calendar days
Circular dependency error Task A depends on B which depends on A Review dependency chain; remove or restructure circular references
Resource over-allocation Same person assigned overlapping tasks Stagger task dates or reassign to balance workload
WBS numbering gaps Tasks created out of order Re-sequence WBS numbers after all tasks are created
Milestone showing duration Duration field not set to zero Update milestone tasks to set duration=0 and milestone=true

Examples

Example 1: Web Application Development

Requirement: "Build a customer self-service portal with account management, ticket submission, and knowledge base search."

Generated Breakdown:

Phase 1: Design (2 weeks)
  1.1 UX Research & Wireframes - 24h - UX Designer
  1.2 Technical Architecture - 16h - Architect
  1.3 API Design - 12h - Tech Lead
  1.4 Design Review - MILESTONE

Phase 2: Development (4 weeks)
  2.1 Account Management Module - 40h - Developer A
  2.2 Ticket Submission Module - 32h - Developer B
  2.3 KB Search Integration - 24h - Developer A
  2.4 UI Implementation - 40h - Frontend Dev
  2.5 Code Complete - MILESTONE

Phase 3: Testing (2 weeks)
  3.1 Integration Testing - 24h - QA Engineer
  3.2 UAT Execution - 16h - BA + Users
  3.3 Performance Testing - 12h - QA Engineer
  3.4 Test Complete - MILESTONE

Phase 4: Deployment (1 week)
  4.1 Production Setup - 8h - DevOps
  4.2 Data Migration - 12h - DBA
  4.3 Go-Live - MILESTONE

Example 2: System Integration Project

Requirement: "Integrate ServiceNow ITSM with Jira for bi-directional incident synchronization."

Generated Breakdown:

Phase 1: Analysis (1 week)
  1.1 Map field mappings SN <-> Jira - 8h - BA
  1.2 Define sync rules and conflict resolution - 8h - Tech Lead
  1.3 Requirements Signed Off - MILESTONE

Phase 2: Build (3 weeks)
  2.1 Configure IntegrationHub spoke - 16h - Developer
  2.2 Build ServiceNow-to-Jira sync flow - 24h - Developer
  2.3 Build Jira-to-SN webhook handler - 24h - Developer
  2.4 Error handling and retry logic - 12h - Developer
  2.5 Integration Complete - MILESTONE

Phase 3: Validate (1 week)
  3.1 End-to-end sync testing - 16h - QA
  3.2 Edge case and error testing - 8h - QA
  3.3 Validation Complete - MILESTONE

Phase 4: Rollout (3 days)
  4.1 Production configuration - 4h - Admin
  4.2 Historical data sync - 8h - Developer
  4.3 Go-Live - MILESTONE

Related Skills

  • spm/cwm-doc-generation - Generate status reports from project data
  • spm/acceptance-criteria - Define acceptance criteria for tasks
  • spm/agile-story-generation - Generate user stories from requirements
  • spm/project-insights - Analyze project performance metrics
  • spm/planning-summarization - Summarize planning data