AI agents for SEO, AI visibility and content ops
Angel Baez builds AI agents that research, write, review, and publish content so companies earn organic search traffic and get cited by AI assistants. At Prepa IN, a Claude managed-agent fleet runs that pipeline daily and grew SEO and AI visibility roughly 20x from a near-zero baseline.
What this niche covers
The work sits at the intersection of three jobs that used to live in separate teams:
- SEO
- Technical foundations and content that search engines can crawl, understand, and rank. That includes clean URLs, structured data, sitemaps, performance, and posts built around real questions.
- AI visibility
- Being cited inside answers from ChatGPT, Perplexity, Google AI Overviews, and similar assistants. That rewards a direct answer near the top, FAQ blocks, sourced facts, and machine-readable discovery surfaces such as llms.txt.
- Content ops
- Daily pipelines that audit the corpus, pick the next topic, draft or update a post, generate assets, fail closed on review, and publish only when a separate reviewer passes. Humans stay in the loop through digests, not by babysitting every step.
How the agent systems work
The production pattern that compounds is a thin coordinator plus specialized sub-agents, each with one job and a fresh context. On Prepa IN's blog fleet that means seven daily-roster agents (SEO auditor, content strategist, writer, updater, thumbnail generator, reviewer, publisher) plus a weekly pipeline meta-auditor.
Deterministic scripts own corpus scoring and queue ranking. Models orchestrate and make the judgment calls that need judgment. Publishing is review-gated: a fail-closed reviewer must leave a passing report on disk, and the publisher refuses to ship without it. Every cycle ends with a digest email so silence itself is a signal.
The content shape matches what answer engines cite: a direct answer in the opening, question-shaped headings, FAQ sections with FAQPage schema, internal links checked against the corpus, and claims that carry sources. Site-level discovery includes RSS and llms.txt.
Case studies on this site
These write-ups are the primary evidence for the niche. Numbers below are only those already published on angel-baez.com.
Claude managed agents at Prepa IN
Architecture of the autonomous fleet that researches, writes, reviews, and publishes Prepa IN's blog. Reported outcome: roughly 20x SEO and AI visibility from a near-zero starting point.
Read the case studyWordPress to Astro migration
Prepa IN's marketing site and blog moved to Astro on Cloudflare in two working days. Six months later, June 2026 was the best organic month on record at 9,473 sessions. Google impressions moved from about 44,000 a month before the migration to 194,000 the month after and 517,000 by March, with average position improving from 7.8 to 5.8. Year-over-year sessions: April +74%, May +82%, June +121%. The post attributes compounding growth after April to the content pipeline built on the new stack.
Read the migration write-upEmDash CMS in production
Cloudflare's open-source, AI-first CMS (EmDash) went live for Prepa IN's blog the day after the developer preview. The stack treats agents as first-class users through a built-in MCP server.
Read the EmDash write-upAylin, an AI teacher for Prepa IN
Aylin is Prepa IN's AI product for personalized learning, built from pedagogy and psychology rather than a generic chatbot. Internal testing began in October 2026, with a planned first version launch in late November.
Read the Aylin announcementRelated AI products
Selected work on the homepage also covers Vera (AI validation of applicant files), Recupera (tax-refund app for salaried employees in Mexico), and Jazmin (WhatsApp AI executive assistant).
See selected work on the homepage
Frequently asked questions
What are AI agents for SEO and content ops?
They are specialized agents that own phases of a content pipeline: auditing the corpus, choosing topics, drafting or updating posts, generating assets, reviewing against a fixed checklist, and publishing when review passes. On Prepa IN, a thin coordinator sequences seven daily agents plus a weekly meta-auditor.
How is AI visibility different from classic SEO?
Classic SEO optimizes for a results page. AI visibility optimizes for being cited inside an answer. Both still need crawlable pages and strong content, but answer engines additionally reward direct openings, FAQ structure, sourced claims, and discovery files such as llms.txt.
Do the agents publish without a human?
Yes, when review passes. Publishing is fail-closed: a separate reviewer agent must leave a passing report on disk, and the publisher will not ship without it. Humans review after the fact through a daily digest and can unpublish anything wrong.
What results has this approach produced at Prepa IN?
The managed-agent fleet write-up reports roughly 20x SEO and AI visibility from a near-zero baseline. Separately, after the WordPress to Astro migration, June 2026 reached 9,473 organic sessions (a site record), and Google impressions rose from about 44,000 a month pre-migration to 517,000 by March, with average position improving from 7.8 to 5.8.
What stack do these systems use?
Prepa IN's content stack described on this site includes Astro on Cloudflare Workers, EmDash CMS with MCP for agents, Claude managed agents (Haiku and Sonnet roles), Cloudflare R2 for the SEO score cache, RSS, and llms.txt. Angel's broader delivery work uses TypeScript, Python, and Node.
Who is this work for?
Teams that need consistent organic growth and AI citations without staffing a full editorial machine. The documented production example is an education company (Prepa IN) whose marketing site is the top of an enrollment funnel.
How do you measure visibility in ChatGPT and similar assistants?
The Prepa IN case study tracks how often the brand appears when students and families ask relevant questions in ChatGPT and other AI assistants, alongside Google Search Console impressions and clicks.
Start a project
Building an agent fleet for SEO, AI visibility, or content ops? Angel takes on a small number of project engagements per quarter. Get in Touch