Programmatic SEO Agent
Generate a real landing page in your voice, not AI slop.
Reads the live SERP for a target query, drafts a landing page sharper than what's ranking, plus 5–8 FAQs from real long-tail queries. One entity or thousands.
Who it's for: Marketing and growth leaders who want programmatic SEO surface area but can't ship hand-written pages fast enough — and don't want generic AI content that ranks worse, not better, under Google's HCU and AI Overviews.
How it works
01 · BRIEF
One row in, one page out
Take a template type (comparison, glossary, location, listicle, vertical landing) plus the entity to instantiate it for. The same template runs across thousands of entities — but each page is written, not stamped.
02 · RESEARCH
Read the live SERP
WebSearches the actual SERP for the target query. Reads what's already ranking. Identifies the real user intent (informational / commercial / navigational / transactional) and the gaps in current results.
03 · WRITE
Outrank by being specific
Drafts the page from the SERP outward — more specific, better structured, more concrete than what ranks. Real entities, real numbers, named examples. No 'in today's fast-paced world.' No 'leveraging.' No filler.
04 · PUBLISH
Ship-ready output
Outputs structured JSON: title, H1, slug, meta description, sections with H2/H3 hierarchy, 5-8 FAQ entries, internal link suggestions, schema.org JSON-LD, CTA. Ready to push to your CMS or render as static HTML.
Signals it can watch
Target keyword + search intent
Template type (comparison, glossary, location, listicle, vertical landing)
Entity / topic to instantiate
Live SERP for the query
Long-tail FAQ queries
What you give it
A page brief
Template type, target entity, primary keyword, search intent, audience, CTA destination. One row of input per page. Iterate at any scale — 1, 100, 10,000 pages.
Your writing voice (as a skill file)
How your brand actually writes. The agent reads this on every run, so the output reads like your team wrote it — not a generic LLM.
Your CTA destinations
Which pages convert (book demo / signup / waitlist / pricing). The agent picks the right CTA for the page's search intent.
What it produces
SEO metadata (title, H1, slug, meta description)
Keyword-leading title within character limits. Readable H1. Slug. 150-160 char meta description. Primary + secondary keywords surfaced as fields, not stuffed.
Structured page body
Intro that answers the query in the first 100 words. H2 sections with optional H3 subsections. Markdown body. Optional comparison table. 5-8 schema.org-compatible FAQ entries pulled from real long-tail queries.
Internal links + schema.org JSON-LD
Suggested internal anchor links to other pages on your site. schema.org JSON-LD blob ready to embed for FAQ / Article / Product / Comparison schema. Word count estimate.
How we use it
Writes the public skill and resource pages on this site.
Every public skill-file page on growthengineer.ai (icp-qualification, founder-outreach, sales-leader-outreach, cold-email-writing, etc.) was drafted by this agent against a brief that pointed at the underlying skill markdown plus the target keyword. The skill files themselves are also published, so the agent's voice is auditable in public.
See the skill files →
Tools used
Web researchCMS / static publish target
Skills it uses
Want this running for your team?
We deploy this agent against your stack, wire up the routing, and turn it on. You keep the keys, the data, and the rep relationships.
Book intro call