How To Automate Real SEO Work With Live SERP Data
This is the best time in the HISTORY of SEO to be doing SEO.
Two Reasons:
Which means ALL the competitors have sub-optimal websites.
And that's the opportunity!
The ONLY people that can keep up with all the new demands are the people using evidence-based SEO automation, so you end up ranking better with a lot less effort.
Most SEO automation fails because the model is guessing.
It gets a URL, a keyword, and a vague instruction like "optimize this page."
Then it gives you the same generic advice everyone has seen a thousand times: match intent, write helpful content, avoid over optimization.
That is NOT what this is.
Instead, this article gives you 17 copy/paste SEO automation recipes that use REAL-TIME SEO intelligence (this type of report)
Except you never see this report because it goes DIRECTLY inside Codex, Claude Code, or another coding agent with the api.on-page.ai MCP connector. (You can connect ANY report you want however as far as I know, this is the first one in the world to provide this entity based, competitor analysis report to AI agents. The better the report, the better the results are.)
The goal is not to mass produce AI content. In fact, it's quite the opposite, it's to preserve human written content as much as possible.
The goal is to make an AI agent do the actual SEO work using REAL SEO data: add missing relevant entities, fix page structure issues, insert relevance internal links within the main content, make controlled edits, rescan the result, and leave an audit trail.
The proof is in the results.
Here is the kind of thing I care about: measurable movement after controlled SEO work, not "AI content" for the sake of AI content.
Good SEO uses the best AI models in the world, it preserves the human writing and it fixes all the other parts that make Google, Bing and ChatGPT fall in love with your content.
The difference between old SEO and the NEW SEO is that you might have been able to do this ONCE in the past, and it would take 8 hours for a single page...
In new SEO era, I'm doing an entire site with hundreds of pages while I'm out at lunch... and it's done just as well, if not better, than it was in the past.
How to use these recipes
1. Connect the On-Page.ai MCP: https://api.on-page.ai
2. Pick the workflow/recipe that matches the job.
3. Paste it into Codex or Claude Code.
4. Replace the example URL, keyword, sitemap, city, or client details.
That's it!
The original full guide is here: https://on-page.ai/pages/automate-seo/
The 17 SEO Automation Recipes
1. Recover a Stuck Page in ONE Command
2. Site Wide Internal Links
3. Single Page Internal Links, Detailed Version
4. Single Page Internal Links, Simple Version
5. Site Wide Refresh for Old or Stale Pages
6. Light Page Refresh, Single Page
7. Standard Optimization, Single Page
8. Standard Optimization, Site Wide
9. Full Client Website Audit PDF
10. Single Page Audit PDF
11. Advanced Page Diagnostic: Why Is This Page Not Ranking?
12. Sub Headline Optimization, Single Page
13. Image and Alt Text Optimization, Single Page
14. Local Page Diagnostic: Why Is This Local Page Not Ranking?
15. Local Page Tuning, Standard Optimization
16. Local Website and GBP Alignment Verification
17. Local Website Cannibalization Checker, City or Region Audit
1. Recover a Stuck Page in ONE Command
Use this when a page is indexed and published, but stuck below where it should be. The agent scans the page, reads the ranking data, finds the gaps, edits carefully, rescans, and produces a before and after audit trail.
2. Site Wide Internal Links
Use this when you want to build internal links across a site without guessing. The workflow finds relevant source pages, chooses natural anchors, inserts links carefully, and records what changed.
3. Single Page Internal Links, Detailed Version
Use this when one important page needs stronger internal links from the most relevant pages on the same site. This version is more detailed and better for priority pages.
4. Single Page Internal Links, Simple Version
Use this when you need a faster, lighter internal linking pass for one target page. It keeps the task focused and avoids turning a small fix into a full site project.
5. Site Wide Refresh for Old or Stale Pages
Use this when a site has old pages that still have value but need fresh research, better coverage, updated examples, internal links, and a clearer audit trail.
Recipe 5 continues below (x has limits on the code block size so this goes after the success part)
6. Light Page Refresh, Single Page
Use this when one page needs a careful refresh without a full rewrite. The goal is to update stale sections while preserving the voice and useful human written content.
7. Standard Optimization, Single Page
Use this when a single page needs a normal optimization pass based on live SERP data. The workflow improves relevance, entities, headings, internal links, and obvious weak spots.
8. Standard Optimization, Site Wide
Use this when you want to run a standard optimization process across a site in a controlled way. It is designed for batches, manifests, checkpoints, and resumable work.
Recipe 8 continues below
9. Full Client Website Audit PDF
Use this when you need a client ready website audit that explains what is wrong, what matters, what to fix first, and how the recommendations connect to real ranking data.
10. Single Page Audit PDF
Use this when you need a focused audit for one page. It is useful for sales calls, client reporting, prioritization, and diagnosing a page before investing in edits.
11. Advanced Page Diagnostic: Why Is This Page Not Ranking?
Use this when a page should be ranking better and you need to know why it is underperforming. This is a diagnostic workflow before you decide what to change.
12. Sub Headline Optimization, Single Page
Use this when the main content is mostly good, but the subheadlines are weak, generic, or missing opportunities to reinforce relevance and user intent.
13. Image and Alt Text Optimization, Single Page
Use this when images and alt text need to support the page instead of sitting there as decorative afterthoughts. The agent improves image context without keyword stuffing.
14. Local Page Diagnostic: Why Is This Local Page Not Ranking?
Use this when a local page is not ranking and you need to compare it against the real local SERP, competitors, service coverage, location signals, and page structure.
Recipe 14 continues below
15. Local Page Tuning, Standard Optimization
Use this when a local page is close, but needs practical tuning based on what is winning in the target city or service area.
16. Local Website and GBP Alignment Verification
Use this when you need to verify that a local website and Google Business Profile are telling the same story. The goal is alignment, trust, and fewer local ranking conflicts.
17. Local Website Cannibalization Checker, City or Region Audit
Use this when a multi location or service area site may have pages competing against each other. The workflow finds city, service, and regional cannibalization problems.
Recipe 17 continues below
And that's my personal recipe book.
Seventeen workflows, ranging from full page recovery, internal linking, full-site refreshes, single-page optimization, competitive analysis, cannibalization audits, all built on live SERP data and all designed to preserve your team's human-written content while producing an audit trail.
But there's one piece missing.
Every recipe above assumes your AI agent has real SERP intelligence data. That's the difference between automation that actually moves rankings and automation that's confidently wrong. Without live scan data, Claude or Codex is guessing (and guessing badly) because their training data is months or years stale.
That's what the On-Page.ai MCP connector solves.
Connect it once. Your agent now sees what's actually ranking, the entities your competitors are covering, the signals the leak documents showed Google cares about, and everything else these recipes need to work. The recipes above run out of the box.
If you have any questions, let me know and I'll make sure to answer everything I can.