---

name: citlyze-action-plan description: Turn Citlyze optimization recommendations into a prioritized, sequenced action plan. Use when asked "what should we do to improve AI visibility", "prioritize the recommendations", "make an AEO/GEO action plan", or after a visibility report surfaces problems. Requires the Citlyze MCP server. license: MIT metadata: author: Citlyze (https://www.citlyze.com)

Citlyze action plan

Convert the recommendation backlog into a two-week plan a marketing or content team can execute.

Prerequisites

The Citlyze MCP server must be connected (setup guide). Tools are read-only; marking recommendations planned or done happens in the app.

Workflow

1. Pull the backlog. Call list_recommendations. Rows arrive highest priority_score first, each with title, summary, category, status, impact, confidence, effort, and target_url.
2. Filter. Drop rows whose status shows they are already done or dismissed; keep new and in-progress work. Treat category as a free-form label: group by whatever values appear.
3. Build an impact/effort matrix. Place each recommendation in one of four quadrants (high impact + low effort first). Where confidence is low, say what evidence would raise it before committing serious effort.
4. Cluster by page. Group recommendations sharing a target_url; one page edit often clears several items at once. Clusters usually beat one-off fixes.
5. Sequence into a plan:
   • Week 1 - quick wins (high impact, low effort) plus anything unblocking other items.
   • Week 2 - the highest-impact medium-effort cluster.
   • Backlog - everything else in priority order, with a one-line reason each.
6. For each planned item state: what to change, on which URL, who plausibly owns it (content, dev, PR), and how success will show up in Citlyze afterwards (better mention rate, new citation, improved visibility on a named engine).

Interpretation rules

• priority_score already blends Citlyze's view of value; use it as the default sort but override with the user's business context when they give any.
• Never promise a score outcome; frame expected results as the metric to watch in the next measurement window.
• If two recommendations conflict on one page, surface the conflict instead of silently picking one.

Learn more: Recommendations in Citlyze - citlyze.com