ai-optimization-planner

name	ai-optimization-planner
alias	/ai-optimization-planner
description	Plan AI Overview readiness improvements — RAG retrievability, E-E-A-T signals, query fan-out coverage, semantic HTML, schema markup, agentic channel setup.
version	1.0.0
disallowedTools	Bash

AI Optimization Planner

JTBD (Jobs To Be Done)

Primary Job

When my site or content exists and I want to improve its eligibility for Google AI Overviews, I want a structured AIO readiness plan, so I can address the specific signal gaps that affect whether my content is retrieved and cited in AI-generated answers.

Secondary Jobs

• When I have good traditional SEO but my content isn't appearing in AI Overviews, I want to understand which AIO-specific signals are missing.
• When I'm preparing an e-commerce or local business site for Google's emerging agentic search experiences (Merchant Center, GBP, UCP), I want a readiness assessment.
• When I know I need to improve E-E-A-T but don't know where to start, I want a prioritized plan with specific improvements across Experience, Expertise, Authoritativeness, and Trustworthiness.

This Skill Is For

• A team with existing content that isn't appearing in AI Overviews despite good traditional SEO.
• A publisher wanting to understand the E-E-A-T signal gaps that reduce AIO inclusion probability.
• A commerce or local business team preparing for Google's agentic search channels.
• An engineering team needing a concrete schema markup and semantic HTML implementation plan.

This Skill Is NOT For

• Writing new content from scratch — use copy-planner for that.
• Structural SEO review (heading hierarchy, meta descriptions, intent alignment, internal linking) — use seo-advisor for that.
• Adding llms.txt, AI-specific markup, or keyword variants for AI modes — Google's AIO guide explicitly states these are not AIO signals.
• Keyword research, SERP analysis, or traffic prediction — use Ahrefs, SEMrush, or Google Search Console.

Resolution Paths

User Situation	What Happens	What They Leave With
Content exists, not in AI Overviews	AIO planner audits E-E-A-T, fan-out, schema, and RAG eligibility	Tiered improvement plan with acceptance criteria
Preparing for agentic channels	Phase 3e assesses Merchant Center, GBP, UCP readiness	Agentic Channel Readiness report with Tier 3 tasks
Schema markup confusion	Phase 3d maps expected vs actual Schema.org types	Specific schema property list to implement
Discovery before planning	Phase 2 surfaces community tools and MCPs	Tooling survey with availability table

When to Escalate

• Structural SEO first: Always invoke seo-advisor before or in parallel — structural SEO is a prerequisite for AIO eligibility, not a substitute.
• Content gaps require original writing: Escalate to copy-planner + copy-critic when Phase 3b (E-E-A-T) identifies MAJOR commodity content gaps.
• HTML/multimedia restructuring: Escalate to web-design-planner or design-partner when Phase 3d identifies MAJOR semantic HTML landmark or layout gaps.
• Accessibility overlap: Escalate to a11y-planner when semantic HTML improvements overlap with accessibility compliance needs.

---

Purpose

Google's AI Overviews are powered by RAG (Retrieval-Augmented Generation). Content that appears in AI Overviews satisfies a different signal set than content that ranks in traditional search:

1. Snippet eligibility — not just indexed, but eligible for Google to extract and use
2. Non-commodity content — first-hand Experience that distinguishes original from recycled material
3. Query fan-out coverage — addressing the related queries Google generates, not just the primary query
4. Semantic HTML and schema markup — structure that browser agents can parse as reliably as humans
5. Agentic readiness — for transactional sites, the structured data and platform integrations that Google's agents need

seo-advisor covers the structural layer (headings, meta, intent, internal linking). This skill covers the AIO-specific layer above it.

---

What This Skill Does

• ✓ Discovers and surveys available MCP tools (Search Console, Playwright, web search) for the planning session
• ✓ Audits indexing status and snippet eligibility (RAG prerequisites)
• ✓ Assesses E-E-A-T across all four dimensions with STRONG/PARTIAL/ABSENT ratings
• ✓ Generates query fan-out analysis (5–8 related queries) and maps content coverage
• ✓ Audits semantic HTML landmark elements and Schema.org markup gaps
• ✓ Assesses multimedia richness (original vs stock, informational vs decorative)
• ✓ Evaluates agentic channel readiness (Merchant Center, GBP, UCP) for transactional sites
• ✓ Produces a tiered implementation plan (Tier 1: quick wins, Tier 2: content investment, Tier 3: platform/architecture)
• ✓ Defines test strategy and review checkpoints with companion skills

What This Skill Does NOT Do

• ✗ Write or rewrite content
• ✗ Audit structural SEO (heading hierarchy, meta descriptions, internal linking)
• ✗ Recommend llms.txt or AI-specific markup (not AIO signals per Google)
• ✗ Predict rankings or traffic
• ✗ Run shell commands, execute code, or query live APIs directly (use MCP tools instead)

---

MCP and Tool Baseline

These tools enhance the plan's accuracy when available. The planner degrades gracefully when they are not available but flags assumptions as FRAGILE.

Tool	Purpose	Impact If Unavailable
Search Console MCP / gsc-cli	Verify indexing status and snippet eligibility	Indexing assumed; flagged FRAGILE
Playwright MCP	Inspect as-rendered DOM, schema markup, landmark elements	Source-only HTML review
Web search MCP (Exa, Brave, Tavily)	Test current AIO appearance; validate fan-out query coverage	Manual testing instructions provided
Lighthouse CLI / PageSpeed Insights API	Core Web Vitals as indirect indexing quality signal	CWV noted as out-of-scope
Schema.org Validator	Validate schema markup before deployment	Manual validation instructions

External Discovery

Phase 2 of the planning protocol uses these discovery mechanisms to surface community tools:

• discover-skills.sh (meta-router repo): Run with keywords "AIO", "Search Console", "E-E-A-T" to find relevant community skills
• GitHub search: Search "Search Console" site:github.com skill SKILL.md and "E-E-A-T" OR "AI Overviews" site:github.com claude skill

Discovery is surfaced as options for the user. External skills are never auto-invoked.

---

Use_When

• Content team reports "our articles have good SEO but don't appear in AI Overviews"
• E-commerce site preparing for Google Shopping AI responses (Merchant Center, Product schema)
• Local business wanting to appear in local agentic responses (GBP, LocalBusiness schema)
• Developer team needs a concrete Schema.org implementation checklist
• Content strategist wants to understand E-E-A-T gaps before commissioning new content
• Pre-publishing AIO readiness check on high-investment content

Do_Not_Use_When

• Need to write content from scratch → use copy-planner
• Need structural SEO review (heading hierarchy, meta, internal links, intent) → use seo-advisor
• Need keyword research or SERP analysis → use Ahrefs, SEMrush, Google Search Console
• Planning to add llms.txt or AI-specific markup → read Google's AIO guide first; these are not signals
• Content doesn't exist yet → start with copy-planner then return here for AIO review

---

Paired With

• ai-optimization-critic: Companion reviewer — invoke at review checkpoints 1 (post-audit) and 3 (post-content-investment)
• seo-advisor: Structural SEO complement — run first or in parallel; invoke at checkpoint 2 (post-Tier-1)
• copy-planner: When E-E-A-T audit identifies MAJOR commodity content gaps requiring original writing
• copy-critic: After Tier 2 content investment, before final ai-optimization-critic review
• web-design-planner / design-partner: When semantic HTML or multimedia gaps require layout/component restructuring
• a11y-planner: When semantic HTML improvements overlap with accessibility compliance

---

Steps

Standard Mode

When invoked with /ai-optimization-planner [site or content URL/samples]:

1. Run Phase 1: Scope and Context — establish content type, agentic goal, risk level
2. Run Phase 2: Discovery and Tooling Survey — MCP availability, companion routing, community skill discovery
3. Run Phase 3a–3e (conditional on content type): RAG retrievability, E-E-A-T, fan-out, semantic HTML, agentic channels
4. Run Phase 4: Assumption Register
5. Run Phase 5: Test Strategy
6. Run Phase 6: Implementation Plan (Tier 1 / Tier 2 / Tier 3)
7. Run Phase 7: Review Checkpoint Plan
8. Output all required headings (see Output Format Contract)

Output Format Contract

These headings must appear in the output (load-bearing for parsers):

## Feature Overview
## Tooling Survey
## RAG Retrievability Report
## E-E-A-T Signal Matrix
## Query Fan-Out Coverage Matrix
## Semantic HTML Signal Map
## Agentic Channel Readiness
## Assumption Register
## Implementation Plan
## Test Strategy
## Review Checkpoint Plan

## Agentic Channel Readiness must be present for all outputs — use "N/A: non-transactional content" for informational sites.

---

Companion_Skills

• ai-optimization-critic: Companion reviewer for checkpoint reviews
• seo-advisor: Structural SEO prerequisite and checkpoint 2 reviewer
• copy-planner: Content investment companion for Tier 2 gaps
• copy-critic: Content quality reviewer before AIO eligibility re-review
• web-design-planner: Platform/architecture companion for Tier 3 semantic HTML work
• a11y-planner: Semantic HTML companion when accessibility overlaps AIO

---

Version: 1.0.0 Last Updated: 2026-05-19 Standalone Ready: Yes Perspective Module: No