Sidecar AI Transformation

Use this skill when the user needs to analyze an AI transformation scenario where core systems, workflows, permissions, compliance, or organizational inertia make direct mainline change difficult.

When To Use

• The customer wants AI, but ERP, CRM, MES, ticketing, approval, or other core systems are hard to change quickly.
• A PoC has value but is blocked by permissions, integration, compliance, responsibility ownership, or unclear ROI.
• The user asks whether to use a Sidecar, external innovation loop, sandbox, plugin-style delivery, or small closed loop before mainline integration.
• The user needs PoC to Beta to production stage-gates, governance, RACI, audit, rollback, or mainline merge criteria.

How To Use

1. Read README.md for the SIDE-CAR seven-step framework and delivery gates.
2. Use prompt.md to run the analysis or to generate a staged prompt for a specific client context.
3. Use checklist.md to validate readiness, risk boundaries, and handoff details.
4. Use evaluation.md to judge whether the output is complete enough for project review.
5. Use workflow.md when the user needs the full discovery-to-governance operating path.

Output Expectations

Produce Markdown suitable for a project memo or solution appendix. Include:

• Sidecar suitability judgment: recommended, cautious, or not recommended.
• Resistance map and stakeholder map.
• Mainline vs Sidecar capability boundary table.
• Interface, data, permission, audit, and rollback design.
• PoC, Beta, production, and mainline-decision stage-gates.
• Commercial case, adoption evidence loop, top risks, and mitigation actions.

Do not invent missing client data. Mark unknown items as pending confirmation.