Last updated: 2026-10-01
Six steps, from your order to the PDF in your inbox. Nothing is skipped, nothing is faked.
Production starts only after we confirm your email and that the correct amount was paid. Orders with a payment mismatch are held for review — we never produce an audit for an unverified order.
We run real buying-intent questions — “Who's the best {service} in {area}?”, “Recommend a reliable {service} near {area}” — across three legs: GPT, Gemini, and live web search. The prompt text is identical on all three legs so the results are comparable. Each prompt runs exactly once: no re-rolls, no cherry-picking. All queries go through official API surfaces — the providers' consumer chat apps are never scraped.
Why this matters: if we mentioned your business in the prompt, the result would be biased. We don't — we ask the way your customers ask.
Every answer is archived byte-for-byte — unedited, unsummarized — with the model version and UTC timestamp, in a per-order evidence file. For each answer we record whether you were mentioned, named, recommended, which competitors were named instead, and which sources were cited.
Your site is graded against 20 checks: 12 machine checks (AI-crawler access in robots.txt, llms.txt presence, JSON-LD completeness, answer-capsule structure, quotable-definition density, statistic density, FAQ coverage, comparison structures, freshness signals, render check, meta robots directives, sitemap health) and 8 manual checks (third-party mention footprint, citation-worthy assets, review and directory presence, competitor citation gap, author credentials, external citations, live prompt tests, internal hub structure). Each check grades PASS, WARN, or FAIL.
A note on the llms.txt check: it grades machine-readability for coding agents — a sitemap-for-bots nicety — not AI visibility. Current research finds AI assistants essentially never fetch llms.txt files, so a missing one is never framed as costing you mentions or citations.
An AI pipeline assembles the report from the evidence — scores, competitor tables, and a fix list ranked by expected impact divided by effort, quick wins first. Then a deterministic QA gate runs before anything ships: zero unfilled placeholders, every finding traced to a dated evidence record, and the score math recomputed independently from the evidence (never copied from a spreadsheet cell). A finding without a trace doesn't go in the report.
A branded PDF lands in your inbox within 48 hours of payment. Every one-time audit includes one revision round for factual errors (wrong hours, wrong service, wrong competitor — we fix it free) and a 30-day money-back guarantee: fewer than three specific, actionable findings means a full refund.
The $299 audit keeps everything in the $99 AI Visibility Audit and adds the complete Google Search side — nine fixed sections, every one cold-auditable with zero logins:
| # | Section | What we measure | Data source |
|---|---|---|---|
| 1 | AI visibility & citations | Mention rate, citation rate, share of voice across GPT, Gemini, live web search — query by query, with verbatim answers | Model APIs (GPT- and Gemini-class), 15–25 buying-intent prompts |
| 2 | Google presence & indexation | Is the site indexed? Brand-query ranking, money-keyword visibility (organic, local pack, AI Overviews) | Google SERP data (US-targeted), site: counts, 10–15 keyword queries |
| 3 | Technical SEO crawl | Broken links, redirect chains, crawl blocks, missing/duplicate titles and descriptions, canonicals, HTTPS, thin content | Self-hosted open-source crawler (full site, typically 20–200 pages) |
| 4 | Core Web Vitals | Lab scores plus real-user field data (LCP, INP, CLS) for the homepage and key pages | Google PageSpeed Insights API (Lighthouse + CrUX field data) |
| 5 | On-page & content quality | Titles, headings, content depth, location relevance, FAQ content — benchmarked against top-ranking competitors | Crawl output + SERP spot-checks vs. competitors |
| 6 | Schema & structured data | All JSON-LD/microdata extracted and validated against Google's documented required and recommended fields | Offline extraction + local validators (no key needed) |
| 7 | Local SEO / Google Business Profile | Listing claimed? Category, hours, photos, review count/rating/recency, NAP character-match vs. your website, citation consistency across major directories | Public GBP data + direct fetches + SERP citation checks |
| 8 | Backlink snapshot (sampled) | Referring domains, authority score, follow/nofollow mix, top-link sample — a labeled snapshot of public data | Moz Link Explorer free + Ahrefs free checker (sampled public data) |
| 9 | Competitive gap analysis | The same checks, lite pass, on 2 named competitors — exactly where they beat you | Same tooling as sections 1–2 |
With about 10 minutes of access — Search Console, Analytics, or your GBP owner account — we can unlock a deeper Phase 2 (real query/CTR data, index coverage, owner insights). We offer it after delivery; it's never required.
One sentence each: $99 tells you what AI says about you. $299 tells you everything about your Google presence too.
| Included | $99 AI Visibility Audit | $299 Full AI + SEO Audit |
|---|---|---|
| AI prompt battery (GPT, Gemini, live web search) | ✓ Full battery, verbatim answers | ✓ Full battery, verbatim answers |
| 20-check AI-visibility scorecard | ✓ | ✓ |
| Competitor comparison in AI answers | ✓ | ✓ |
| Prioritized fix list (quick wins first) | ✓ | ✓ |
| Google indexation & rankings (§2) | — | ✓ |
| Technical SEO crawl (§3) | — | ✓ |
| Core Web Vitals, lab + field (§4) | — | ✓ |
| On-page & content quality (§5) | — | ✓ |
| Schema / structured data (§6) | — | ✓ |
| Local SEO / GBP review (§7) | — | ✓ |
| Backlink snapshot, labeled sampled (§8) | — | ✓ |
| 2-competitor gap analysis (§9) | — | ✓ |
| Delivery | Branded PDF, 48 hours | Branded PDF, 48 hours |
| Guarantee | 30-day money-back | 30-day money-back |
Full dated price comparison against the market on the pricing page.
One formula, applied identically to every audit. No black box.
Each engine — GPT, Gemini, live web search — gets its own score out of 100:
Your final score is the rounded mean of the three engine scores. The verdict bands:
The scorecard is our internal rubric, published here — not a Google metric, not an AI-vendor metric. The fix-list ranking (expected impact ÷ effort) is driven by the evidence weights from current AI-search research: crawler access, quotable-definition density, statistic density, third-party mention footprint, the competitor citation gap, and the live prompt tests.
These are non-negotiable — they're what make a paid audit defensible:
An honest methodology includes what it can't do:
The questions owners ask about how the sausage is made.
Because that's what we actually test. We query GPT and Gemini model families and live web search through official API surfaces — not the ChatGPT or Perplexity consumer products. The results closely approximate what users of those products see for the same buying-intent queries, but they aren't identical, and we'd rather underclaim than overclaim.
Real. Every quoted AI answer is captured verbatim — byte-for-byte — from a live API call, archived with the model version and UTC timestamp in an evidence file. We never reconstruct findings, infer model behavior, or present paraphrase as quotation. Each finding in your report cites the exact prompt and date that produced it.
Because free backlink sources only expose a sample of the link graph. We report domain-level authority plus a top-link sample from Moz and Ahrefs free data, and we label it as a snapshot. We never present third-party authority scores as Google metrics. A deep link audit is available as a follow-up once you grant access to your own Search Console data.
No. Every audit is run from the outside, exactly as a customer or an AI assistant would see your business: public web searches, AI-assistant prompts, and your public website. After delivery we can offer a deeper Phase 2 — with about 10 minutes of access to Search Console, Analytics, or your Google Business Profile owner account — but it's never required.
Every audit includes one revision round for factual errors, free. If we stated something wrong about your business — wrong hours, wrong service, wrong competitor — we fix it. Scope stays fixed to the published sections; new checks beyond them are follow-up work, not revisions.
Each engine (GPT, Gemini, live web search) gets its own score out of 100: prompt visibility counts for 60 points (across two core buying-intent prompts: your business named-and-recommended scores 30, named-but-not-recommended 20, mentioned in passing 10, absent 0) and our 20-check site scorecard counts for 40 (each check: PASS 2, WARN 1, FAIL 0). Your final score is the rounded mean of the three engine scores. 70+ means Visible, 40–69 Patchy, below 40 Invisible.
No — and no honest vendor can promise that. We measure exactly where you stand today (mentioned? cited? facts correct?), where competitors beat you, and which fixes will most improve your odds, ranked by expected impact divided by effort. The report describes what the models said, not what they'll say next month.
You receive a branded PDF by email within 48 hours of payment. The data collection itself runs 15–30 minutes, mostly unattended; the rest is synthesis and our pre-delivery QA gate, where every finding is checked against the archived evidence before anything ships.
The $99 AI Visibility Audit: 15–25 real buying-intent prompts across GPT, Gemini, and live web search, verbatim answers, the 20-check scorecard, competitor comparison, and a prioritized fix list. Branded PDF in 48 hours. 30-day money-back guarantee.
Get my audit — $99