What is AI search visibility (AEO/GEO)?

Answer Engine Optimization (AEO) / Generative Engine Optimization (GEO) is getting your brand cited inside AI answers (ChatGPT, Claude, Gemini, Perplexity, and Google AI Overviews), not only ranking in blue links.

Does schema markup get me cited by AI?

Schema and llms.txt are hygiene: necessary so engines can parse you, but not a citation lever on their own. The real levers are brand mentions and Bing ranking. We work all of it, and every change is human-reviewed.

How much does a Prompt Goblin audit cost?

Goblin Scout is a monthly visibility audit at $997/mo. Warband is $3,500/mo and Warlord $9,500/mo. Monthly, cancel anytime, no credits, no meter.

Is there a money-back guarantee?

Yes: a 100% money-back guarantee on the work, not on the algorithm. If we don't deliver your audit, or you're not happy with it within 14 days of delivery, you get a full refund, no argument. We can't promise a specific citation number (AI citation share is volatile and partly outside anyone's control), so we never sell one. We guarantee the work, and we measure the result honestly.

./methodology

Exactly what the scan measures, and exactly what we will not promise.

Prompt Goblin is an AEO/GEO, technical SEO, and accessibility workflow. It measures crawlability, answer-engine citation gaps, static accessibility risks, and human-reviewed fix work. It does not promise a citation number.

four layers

The scan stack

Technical hygiene

Crawl path, robots and llms.txt presence, indexability, head tags, canonical tags, and static Core Web Vitals proxies.

A hygiene issue means a crawler or answer engine may have trouble reading, summarizing, or trusting the page shell.

Static fetches have a blind spot: an unreadable, JS-rendered, SPA, or WAF-blocked site is flagged, never scored 0.

Schema

JSON-LD types present, parseable, and appropriate for the business model.

A schema issue means the page is harder to parse as an entity or service.

Schema is hygiene, not a citation lever. It helps engines parse you; it does not earn a citation.

AI visibility

Answer-engine citation gaps, competitor citation graph differences, freshness, topical depth, and third-party platform presence.

A visibility issue means the answer surface currently points elsewhere or lacks a verifiable source pattern we can work on.

Citation movement is measured over re-runs. No citation number is guaranteed.

Accessibility

Static WCAG 2.1 AA and Section 508 pre-screen signals such as labels, image alt text, language, and landmarks.

An accessibility issue means the page may block assistive technology or fail compliance review.

Automated checks catch only part of WCAG; full conformance needs rendered axe checks and human review.

finding meaning

Severity is a work queue, not theater

HIGH	A live blocker, noindex, unreadable surface, or missing critical entity path that needs human review first.
MED	A fixable issue that can materially improve crawlability, clarity, accessibility, or measurement quality.
LOW	A cleanup or supporting signal that improves the surface without being a standalone citation lever.

recon

The target is inferred, then reviewed

Recon infers category, topic, summary, early stack fingerprints, and up to two competitors when the operator did not provide competitors.
Auto-discovered competitors are flagged research inferences the client confirms, never asserted facts.
The first homepage fetch is cached so later audit nodes do not repeatedly probe the same site.

human gate

Nothing auto-ships

Every recommendation halts for engineer review before a client sees it or a change reaches production. Reports can include code snippets, community drafts, and AI prompts, but the pipeline never posts, sends, deploys, or claims a pass without the matching evidence.

honest broker

The limits are part of the product

We never fabricate metrics, clients, testimonials, or citations.
Schema and llms.txt are hygiene, never a promised citation lever.
Unreadable, JS-rendered, SPA, or WAF-blocked pages are flagged, never scored 0.
Service and government sites correctly use Service, Offer, and OfferCatalog patterns; we do not prescribe commerce schema for them.
Every recommended change is human-reviewed and human-gated. Nothing auto-deploys or auto-sends.
Mock, sample, and demo paths are illustrative and are never reported as real passes.
The refund covers the work, never a citation number.

run the scan ->