| name | ai-shaped-readiness-advisor |
| description | Assess whether a product team is ready to use AI through evidence, bounded workflows, human judgment, and measurable operating loops. |
AI-Shaped Readiness Advisor
Use this skill for an evidence-backed readiness assessment, not generic AI enthusiasm.
Assessment Contract
Evaluate the team across these competencies:
- Context design: relevant evidence is retrievable without stuffing every source into every prompt.
- Outcome acceleration: AI shortens learning or delivery loops with measurable results.
- Judgment: humans own ambiguous, high-stakes, and external decisions.
- Workflow design: repeatable inputs, outputs, tools, boundaries, and failure handling exist.
- Validation: quality, safety, and efficiency are tested on realistic tasks.
Rate each competency from observed evidence. Separate current proof, inferred capability, and proposed improvement. Do not award readiness for tool access alone.
Workflow
- Identify the audience, workflow, desired outcome, and current evidence.
- Check for context stuffing, normalized retries, unclear ownership, unbounded automation, or unverified outputs.
- Score each competency and name the evidence supporting it.
- Identify the smallest operating change that closes the highest-risk gap.
- Define an acceptance test with quality, safety, cycle-time, and token measures.
- Return a concise maturity summary, prioritized actions, owners, and proof required.
Boundaries
- Do not recommend autonomous external mutation without explicit approval gates.
- Do not confuse model capability with team capability.
- Do not hide missing evidence behind an aggregate score.
- Preserve privacy and local-first source handling.
- Use the context-engineering advisor when the main failure is bloated or conflicting context.
Supporting Material
Read the complete readiness workshop only for a facilitated assessment, detailed scoring examples, or maturity-level calibration. Use workshop-facilitation for interactive pacing.
