SEO Recommendations

How the SEO Agent turns keyword gaps into prioritized, copy-paste-ready recommendations you can act on in your daily 5-minute review.

The SEO Agent continuously compares your keyword opportunities against your existing pages, then writes prioritized recommendations that tell you exactly what to change β€” and hands you a ready-to-use prompt to make that change in your coding agent. It skips weak opportunities rather than filling your queue with busywork, so what you see is worth acting on.

Recommendation Types

Each recommendation targets one page and one keyword gap. The types the agent writes are:

  • optimize_title β€” the page title doesn't lead with the right keyword or isn't compelling enough for the search result
  • optimize_meta β€” the meta description needs a rewrite for click-through or relevance
  • optimize_h1 β€” the primary H1 is misaligned with the target keyword
  • optimize_content β€” the page body needs new or restructured content to match search intent
  • new_page β€” no existing page targets this keyword; a new page should be created
  • rebind β€” the keyword is currently pointing to the wrong canonical page; it should be reassigned to a better one
  • technical_issue β€” a crawl or structural problem (missing tag, duplicate title, etc.) is limiting the page's performance

Every recommendation includes a plain-English reason explaining why the agent flagged it, a priority, and a coding-agent prompt you can paste directly into Cursor or Claude Code to implement the change without leaving your editor.

How It Works

  1. The agent scores keyword-page alignment using GSC performance data, search metrics, and crawl results. Recommendations flow into your queue only when the signal is strong enough to justify the change.
  2. Open the Recommendations tab in your Growth Autopilot dashboard. Each pending item shows the type, the target page, the keyword, the priority, and the reason.
  3. Read the reason and decide: Apply or Dismiss. Applying a recommendation binds the keyword's primary page mapping and propagates any secondary keywords to that page. For a new_page recommendation, VibeCom's blog persona picks it up on the next content-generation run and drafts a post for you to review and approve.
  4. For recommendations that require a markup or content change (title, meta, H1, content), copy the attached coding-agent prompt and paste it into Cursor or Claude Code. The prompt is self-contained β€” no additional context needed.
  5. You can also apply or dismiss recommendations directly from your IDE via MCP, without opening the dashboard at all.

Re-Evaluation After 28 Days

Roughly 28 days after a recommendation is applied, the agent re-checks the page against Google Search Console metrics. If the baseline didn't improve β€” clicks, impressions, or position β€” the agent writes a fresh recommendation based on current data. This keeps the queue honest and closes the loop between implementation and outcome. VibeCom does not claim to control or accelerate Google's indexing of your pages; the 28-day window reflects normal GSC data latency.

Next Steps

  • β€” apply and dismiss from your editor
  • β€” the data behind recommendations
SEO Recommendations | VibeCom Docs