Проектирование AI-воркфлоу для маркетинга

★ 6.9 · marketing

ai-research-workflow is a Claude Code skill that designs and documents an AI research workflow from scratch, producing a structured document with contextual analysis, actionable recommendations, an implementation plan, and a risk assessment. Before generating output, the skill reads project context from `memory.md` and `knowledge-base.md`, then moves through four steps: context gathering, framework application (AI Readiness Assessment, Automation ROI Calculator, Human-in-the-Loop Design, RAG Architecture, Agent Orchestration Patterns), document generation, and a quality-validation checklist. Success metrics — Time Saved Per Task, Automation Rate, Error Reduction %, and Cost Per AI Operation — are captured in a table with baseline and target values. The final deliverable includes an executive summary, a risk matrix, and a next-steps checklist with owners and timelines. It suits marketing teams that need a repeatable, KPI-driven AI research process with built-in human review for critical outputs.