AI Workflow Templates
ConsultingintermediateHITL requiredupdated 2026-08-11

SOW Generator (Lite)

Draft a consulting-grade Statement of Work from structured operational intelligence. Facts and labeled assumptions only. Never invent pricing.

#sow#consulting#hitl#markdown
View source →

Tools in the pipeline

  • api
    intel-loader
    Load operating intelligence for the engagement.
  • llm
    sow-model
    Emit structured JSON plus polished Markdown.
  • human
    partner-review
    Sign off scope, commercial placeholders, and unanswered questions.
Stack: LLM · Markdown · JSON

What it does

Generates a Statement of Work from supplied operational intelligence. Separates current state, objectives, scope, out of scope, requirements, deliverables, milestones, dependencies, assumptions, risks, acceptance criteria, success metrics, governance, timeline, staffing, and commercial assumptions. Listed as a free ARC Labs utility in the Master Blueprint (SOW Generator Lite).

Pipeline

  1. Load — structured intelligence + engagement goal.
  2. Draft JSON — every major section as fields.
  3. Draft Markdown — consulting-grade prose from the same fields.
  4. Flag — unanswered questions and pricing placeholders.
  5. Review — partner edits before it leaves the building.

Prompt skeleton

Generate a consulting-grade Statement of Work from the supplied structured operational intelligence.

Rules:
- Use only supplied facts and explicitly labeled assumptions.
- Separate current state, objectives, scope, out of scope, requirements,
  deliverables, milestones, dependencies, assumptions, risks, acceptance
  criteria, success metrics, governance, timeline, staffing, and commercial assumptions.
- Link major scope items to evidence or identified operational needs.
- Identify unanswered questions.
- Never invent pricing. Where pricing is missing, provide placeholders and estimation logic.
- Return: 1) structured JSON  2) polished Markdown.

Guardrails

Never invent pricing or confidential client facts. Label assumptions. Link scope to evidence. Human signs the SOW; the model does not send it.

Structural dry-run

No model call. Validates required fields only.

Prefills the example payload for AI readiness SOW. Edit it, then run.

Related