AI-Assisted SEO Automation (75–80% Coverage)

Overview

An AI-driven workflow now handles the majority of routine on-page SEO tasks — meta titles, meta descriptions, schema injection, and target keyword assignment — reducing what once took a full day of manual work to a matter of minutes. The remaining 20–25% of tasks, primarily heading structure (H1/H2/H3), cannot be safely automated and must be handled by developers at build time.

What the AI Tool Automates

Task	Automatable?	Notes
Meta titles	✅ Yes	Generates keyword-optimized titles at correct length
Meta descriptions	✅ Yes	Writes page-specific descriptions within 160-char limit
Target keyword assignment	✅ Yes	Sets SEO plugin target keyword per page
Schema injection	✅ Yes	Deployed via must-use (MU) plugin; does not break pages
Alt text for standard images	✅ Yes	Skips icons/PDFs that cannot accept alt attributes
H1 / H2 / H3 headings	❌ No	Requires developer; AI manipulation of Elementor breaks sites

How It Works

The tool reads credentials from a central sheet, logs into each WordPress site, crawls all pages via the REST API, audits the active SEO plugin (e.g., Yoast, SEOPress), and then writes optimized values back — all without touching the page builder or theme layer. It skips utility pages (thank-you pages, test pages, etc.) automatically.

Demonstrated result on Reynolds: 27 missing meta titles and 7 missing meta descriptions fixed across 56 pages in under two minutes.

What Still Requires Manual / Developer Work

H1 / H2 / H3 Headings

Heading tags are the one SEO element the AI cannot safely set. Attempts to manipulate Elementor components via the API have caused site breakage requiring restores. Until a safe injection method is identified, headings must be authored by developers at build time.

Process fix adopted: Developer sitemaps must now include the target keyword for each page. With that information, developers can correctly place the keyword in H1, H2, and H3 tags during initial build — eliminating the need for retroactive fixes.

"If they put the target keyword in the H1, H2, and H3 in there, we can do everything else. That's all we really need the developers to do." — Mark Hope

Schema (Complex Cases)

Standard schema (local business, moving company, etc.) can be injected via MU plugin. More complex or site-specific schema may still require review. See ^[1] for implementation notes.

Client Examples

^[2] — AI fixed 27 missing titles and 7 missing descriptions in minutes; Ahrefs health score reached 100. 23 pages still missing H1 tags (manual developer work pending). Google Ads and conversion tracking audits run concurrently; all recommendations implemented.
^[3] — Ahrefs score raised from 60s to 100 after image resize and manual H1 additions. No target keywords were set on any page at time of audit — a gap the new sitemap process is designed to prevent.

Known Limitations & Risks

Elementor breakage — Do not attempt to set headings or manipulate layout components via API. Restore from backup if this occurs.
SEO plugin conflicts — When an MU plugin and an SEO plugin (e.g., SEOPress) both attempt to write schema, conflicts can arise. Audit for duplication after deployment.
Icons and non-image files — Alt text automation skips files stored in the image directory that are not true images (icons, PDFs). These 42-image gaps (as seen on Reynolds) are expected behavior, not tool failures.
Shopify sites — The REST API approach does not apply. Shopify sites require separate handling; Cloudflare bot-blocking can also prevent crawl audits entirely.

Open Research

Mark Hope is investigating safe methods to automate H1/H2/H3 injection without breaking Elementor-based layouts. No reliable approach found as of 2026-03-17.
Evaluating whether SEO plugins (Yoast, SEOPress) can be retired in favor of fully manual/API-driven SEO management to reduce plugin overhead and site fragility.

^[1]
^[4]
^[5]
^[6]
^[7]