(Sofia · Ext. 105 · knowledge base)

Search Visibility Lead.

Keeps you ranked on Google and cited by ChatGPT, Gemini, and Perplexity — audits, fixes, and weekly proof, nothing published without your sign-off.

Launch this role

What it does from day one

(no ramp-up, no aspiration)

Technical SEO crawl (up to ~10-50k URLs) with an LLM-ranked prioritized fix list: broken links, duplicate titles, redirect chains, thin pages
AI-visibility baseline: 25-100 buyer-intent prompts fired at ChatGPT/Gemini/Perplexity/Claude, recording mentions, citations, and competitor share-of-voice
Schema/JSON-LD generation and validation per page, delivered as snippets or staged CMS changes
Answer-first on-page rewrites: titles, headings, 40-60-word direct-answer blocks, FAQ sections — as tracked-changes drafts
Weekly rank + AI-mention tracking with delta reports (DataForSEO + Search Console + rerun prompt panel)
llms.txt generation, robots.txt AI-crawler hygiene, and GSC URL inspection/submission

How the work flows

(brief to logged output)

Onboard & baseline — GSC/GA4/CMS access; full crawl, GEO prompt scan, 16 months of GSC data → audit + AI-visibility scorecard
Plan — audit becomes a prioritized backlog (technical > schema > answer blocks > content > citation outreach); client approves the sprint
Execute — schema, rewrites, new FAQ/comparison drafts, outreach emails — everything lands as a draft, PR, or staged CMS change
Human gate & publish — approval, then push via WordPress/Webflow API and resubmit changed URLs for indexing
Measure & log — weekly re-pulls of ranks, GSC metrics, and AI mentions; delta report written and wins/losses feed the next sprint

The stack it plugs into

(real tools, official APIs)

Google Search Console API — query/page performance, sitemaps, URL inspection (free, 500 req/day)
DataForSEO API — SERP tracking and keyword data at pay-as-you-go pennies vs $500/mo Semrush/Ahrefs API plans
LLM provider APIs (OpenAI, Gemini, Anthropic, Perplexity Sonar) — the GEO prompt panel; Perplexity returns live citations
WordPress REST / Webflow / Shopify Admin APIs — stage and publish meta, schema, and content edits
GA4 Data API — organic sessions and conversion attribution for reporting
PageSpeed Insights API — Core Web Vitals without running Lighthouse locally
Looker Studio + Slack/email — dashboards, approvals, weekly reports

Its machine

(8GB RAM · 4 vCores, dedicated)

An async Python crawler (a few hundred MB for tens of thousands of URLs), one headless Chromium for JS-rendered pages and validation, a scheduler for weekly rank pulls and GEO scans, and a small time-series DB for ranks/mentions. I/O-bound work; 8GB is comfortable with one browser, ≤3-4 pages, and restarts every few hundred pages — only a full JS render of a 100k-page SPA needs chunking.

Seat economics

medium token load

GEO prompt panels are cheap (1,600 short completions/month ≈ a few dollars); content rewrites are the real token cost — a realistic month lands around $50-150 of model spend, well inside the $499 seat, and the bigger line item is third-party API spend, not tokens.

Flat seat from $499/agent/mo — one agent, one dedicated machine. See pricing.

Judged by

(the numbers that matter)

Organic traffic and organic-conversion trend (baseline conversion ~2.7-3%, B2B service pages 2-5%)
Rankings/CTR on target keywords (position 1 averages ~27.6% CTR, falling as AI Overviews absorb clicks)
AI share-of-voice — % of tracked prompts where the brand is cited per engine, vs competitors

Tasks in this role

(each documented in the task library)

Technical SEO Audits
A crawl of up to ~10-50k URLs that comes back as a prioritized fix list, not a spreadsheet dump.
AI Search Visibility
Are ChatGPT, Gemini, and Perplexity recommending you or your competitor? The AI employee fires 25-100 buyer-intent prompts at the major AI engines, records mentions, citations, and competitor share-of-voice, and reruns the panel weekly to prove movement..
Schema Markup Generation
JSON-LD structured data generated and validated per page, delivered as snippets or staged CMS changes — never pushed live without sign-off.
SEO Content Optimization
Answer-first on-page rewrites built for both Google and AI engines: titles, headings, 40-60-word direct-answer blocks, and FAQ sections.
SEO Rank Tracking Reports
Weekly rank and AI-mention tracking with a written delta report: what moved, why, and what happens next sprint.

Never alone

(guardrails — how trust works here)

No unsupervised publishing or site-code changes — every edit, schema injection, redirect, or robots change ships staged for approval (a bad noindex can destroy traffic; Google's scaled-content enforcement has cut sites 50-80%)
No outbound outreach or third-party posting without sign-off — agent drafts link-building emails and directory submissions, humans send
Factual-claims gate on YMYL/product content: stats, health/finance/legal claims, pricing, and competitor comparisons carry sources and pass human review
Escalate anomalies, don't remediate: >20% weekly traffic drop, manual actions, deindexing → alert a human; never attempt disavows or mass deletions autonomously

Full policy: our AI disclosure.

Try it

(alpha)

"AI Visibility X-Ray": visitor enters a domain; in ~90 seconds the agent generates 10 buyer-intent prompts, fires them at ChatGPT/Gemini/Perplexity with answers streaming split-screen — highlighting where the brand is missing and which competitor got cited instead — while a 20-point AEO spot-check runs on the homepage. Ends with a scorecard ('invisible in 7 of 10 buying questions') and a sample JSON-LD fix. The 'watch AI ask ChatGPT about you and come back empty' moment is this role's version of hearing the cold-caller speak. Cold calling ships first — try that demo today.