Summary
This skill equips an agent to leverage generative AI for tactical SEO execution—keyword research acceleration, content creation support, technical markup generation, and SEO automation—while knowing exactly where AI fails and what requires human verification. Invoke this when you need to move fast on repetitive SEO tasks but don't have access to commercial keyword research tools or want to augment existing workflows with AI-assisted analysis and code generation.
SKILL.MD
Execute AI-accelerated SEO workflows
When to Activate
Load this skill when you need to:
- Accelerate keyword research and intent analysis
- Generate or optimize SEO content elements at scale
- Implement technical SEO markup (schema, hreflang)
- Analyze SEO performance data or build reporting visualizations
- Troubleshoot regex, spreadsheet formulas, or Python scripts for SEO tasks
Core Knowledge
AI Capabilities and Limitations
What generative AI does well:
- Summarizing information from multiple sources
- Pattern recognition across large text datasets
- Generating variations of short-form content
- Writing and debugging code (regex, Python, spreadsheet formulas)
Critical limitations:
- Cannot perform accurate mathematical calculations (will hallucinate keyword volume/difficulty numbers)
- Lacks "bigger picture" strategic thinking
- Won't notify you when it misunderstands prompts or lacks necessary data
- Produces confident-sounding output even when wrong
Implication: Use AI for execution speed and ideation, not for strategic decisions or data that requires accuracy. Always verify outputs, especially numerical data.
Constraints / Hard Rules
- Never trust AI-generated keyword metrics. ChatGPT and similar models will fabricate search volume and difficulty scores. Use actual keyword research tools (Ahrefs Keywords Explorer, etc.) for metrics.
- Do not use AI to mass-produce thin content. Google penalizes low-quality content regardless of how it's created. AI makes it easy to create bad content at scale, which increases manual penalty risk.
- Treat all AI outputs with skepticism. Verify factual claims, check generated code before deploying, and review content for accuracy.
- Do not rely on AI for strategic planning. Use it for tactical execution, not for determining overall SEO strategy for a specific business.
Workflow
Keyword Research Acceleration
Generate seed keywords:
- Identify your starting topic
- Prompt AI: "Suggest seed keywords related to [TOPIC], including sub-topics, questions, and similar concepts"
- Take the list and input into Keywords Explorer (or equivalent tool with real data)
- Review actual volume and difficulty metrics
- Navigate to Matching Terms or Related Terms tabs to expand keyword list
Alternative: Use Keywords Explorer's built-in AI feature to suggest subtopics and niche areas directly within the tool.
Analyze SERP intent:
- Search your target keyword and copy page titles from top-ranking results
- Prompt AI: "Categorize these page titles by search intent type: [paste titles]"
- Refine the categorization if needed by asking for specific groupings
- Identify the dominant intent and approximate traffic distribution
More efficient alternative: Use Ahrefs "Identify intents" feature in SERP overview, which shows percentage of traffic each intent receives.
Decision point: Target the intent that captures the majority of traffic. In the article's "LLM" example, 82% of traffic went to definitional content, making that the correct intent to target.
Content Creation Support
Generate title variations:
- Draft your content or outline key points
- Paste content into AI with prompt: "Suggest [number] title ideas for this content: [paste content]"
- Extract useful words, phrases, or angles (don't use verbatim)
- Combine AI suggestions with your own judgment to craft final title
Grammar checking:
- Paste content paragraph or section into AI grammar checker
- Review flagged issues
- Accept, modify, or reject each suggestion based on context
Clean up transcripts:
- Generate raw transcript from interview or video (use any transcription tool)
- Prompt AI: "Clean up this transcript: fix typos, capitalize brand names correctly, add proper punctuation: [paste transcript]"
- Review edited version for quote extraction or article repurposing
Content Optimization
Identify missing topics:
- Use Ahrefs Content Grader (or similar tool) to compare your content against top-ranking articles
- Review automatically identified topic gaps
- For each gap, request explanation: "How should I address [topic gap]? Show example from competing content"
- Decide which gaps are worth addressing based on relevance to user intent
Generate meta descriptions at scale:
- For each page, note key content elements
- Use AI meta description generator with inputs: page content summary, desired tone, number of variations
- Review outputs and select strongest option (or combine elements)
Note: Google frequently rewrites meta descriptions, so speed matters more than perfection here.
Evaluate against helpful content guidelines:
- Use a specialized GPT trained on Google's helpful content guidelines (like Aleyda Solis's)
- Submit your content and a competing article
- Review automated feedback for obvious issues
- Apply professional SEO judgment to determine which suggestions to implement
Technical SEO Implementation
Create schema markup:
- Identify content type needing schema (recipe, review, product, etc.)
- Prompt AI: "Generate [schema type] markup for: [paste content details]"
- Review output and add any missing required fields (author, dates, etc.)
- Validate using Google's Rich Results Test before deploying
Generate hreflang tags:
- List all language/region versions of your page and their URLs
- Prompt AI: "Generate hreflang tags for these page versions: [paste list]"
- Review output for accuracy
- Implement in page
<head>or sitemap
Analysis and Reporting
Construct regex queries:
- Describe your data pattern matching need in plain language
- Prompt AI: "Write a regex query that [describe goal]. The data structure is: [describe structure]"
- Test the regex on sample data
- If errors occur, share error message with AI for troubleshooting
Example use cases from article:
- Extract URLs from email addresses
- Filter URLs by crawl depth
- Filter localized content by country code in URL path
Build spreadsheet formulas:
- Describe your spreadsheet structure (column names, data types)
- Describe the calculation or filter you need in plain language
- Prompt AI: "Create a Google Sheets formula that [describe goal]. The spreadsheet structure is: [describe structure]"
- Test formula on sample data
- Share error messages with AI if formula breaks
Write Python scripts:
- Describe the automation task in detail
- Prompt AI: "Write a Python script that [describe goal]. Include comments explaining each section."
- Review code for security issues (especially for web scraping)
- Test on small dataset before scaling
- Share error messages with AI for debugging
Example use cases from article:
- Web scraping for storing webpage data
- API calls to collect bulk traffic and backlink data
Create data visualizations:
- Export data from Ahrefs (or other SEO tool) to CSV
- Upload to ChatGPT or similar AI with visualization capabilities
- Prompt: "Create a [chart type] showing [specific metric] over time, highlighting [specific patterns]"
- Refine visualization with follow-up prompts
Example visualizations from article:
- Organic traffic over time with anomaly highlighting (Google updates)
- Desktop vs mobile rankings comparison
- Seasonal patterns in backlink acquisition
Output Contract
When executing this skill, you produce:
For keyword research:
- Expanded seed keyword lists (verified with actual metrics from keyword tools)
- SERP intent categorizations with traffic distribution percentages
- Recommended content type based on dominant intent
For content creation:
- Multiple title/header variations with useful phrases extracted
- Grammar-checked content with flagged issues
- Cleaned transcripts ready for quote extraction
For content optimization:
- Topic gap analysis with implementation recommendations
- Meta descriptions in specified tone (multiple variations)
- Helpful content assessment with specific improvement areas
For technical SEO:
- Valid schema markup (JSON-LD format, tested in Rich Results Test)
- Complete hreflang tag sets for multi-language/region sites
For analysis:
- Working regex queries for pattern matching
- Functional spreadsheet formulas for data manipulation
- Debugged Python scripts for SEO automation
- Publication-ready data visualizations
All outputs should include your verification notes highlighting what was checked and what limitations remain.