First, the honest part: this is a sample, not a client report
A diagnostic report exposes a client's real weaknesses — we never showcase a real one. Every figure here comes from Negency's published flagship study, the AI Search Visibility Report for Chinese Companies Going Global (2026, first edition): 30 Chinese exporters' global website home domains; Negency AI visibility audit engine, around 30 signals, five categories; single-day sample, 30 October 2026. The report measures website technical visibility only, not brand mention rates in AI answers. The "median profile" scorecard below follows part one of a real diagnostic report exactly — for your company, every number comes from measuring your own domain page by page.
Sample 1: the scorecard (median profile, 30-company benchmark)
Total: 80 / 100 (median of 30; mean 80.1; range 55–98; 7 sites above 90, 6 below 70)
Crawlability: 91.1 (mean of 30) — can AI crawlers get in and get around? 21 sites at full marks, the strongest dimension
Performance: 87.3 — do load and rendering slow down crawlers and visitors?
Trust signals: 82.1 — are entity facts, contacts and third-party profiles consistent and checkable?
Entity structure: 71.1 — can a machine read, unambiguously, who you are, what you sell, who it is for?
Machine readability: 69.2 (mean; median 68.5) — structured data, llms files, semantic markup; the weakest dimension, only 2 sites at full marks
A scorecard shows the bottleneck dimension — not a total to celebrate or mourn. The median profile says most companies spend in the wrong place: plenty of content, while machine readability and entity structure drag. Two statistics matter: median 80 and mean 80.1 sit close, so the distribution is roughly symmetric; machine readability's mean (69.2) and median (68.5) are close too — the weakness is broad.
Sample 2: what an itemised finding looks like
Every finding carries four things: problem, evidence, impact, fix. The three most frequent across the 30 companies, in the real format:
Problem: no JSON-LD structured data anywhere on the site (17 of 30). Evidence: source-level check of the homepage and core product pages — no Organization, FAQPage or Product markup. Impact: engines guess your entity and product facts from prose — guesses drift descriptions and lose citations. Fix: add Organization plus the four core markup types, word-consistent with visible text. Specification: Schema.org Organization and Google's structured-data introduction.
Problem: H1 missing or not unique (13 of 30). Evidence: the crawl inventory lists every page's H1 text. Impact: several topics compete on one page; engines cannot tell which passage to quote — Q&A content suffers most. Fix: one H1 per page, aligned with the question the page answers.
Problem: sitemap problems (15 of 30). Evidence: item-by-item check of sitemap reachability and contents — dead addresses submitted, core pages omitted, conflicts with robots rules. Impact: crawlers take wrong turns or miss key pages; new content enters the index slowly. Fix: rebuild the sitemap against the real page inventory; confirm robots is not blocking the sitemap itself.
Sample 3: the 90-day roadmap skeleton
Month 1 · Technical and entity layer: clear crawl blockers, add structured data, unify facts and profiles — get machines reading you right first. Acceptance: audit scores rise; AI crawler visits and success rates rise in server logs.
Month 2 · Content and Q&A layer: publish question-shaped content from buyer questions; category definition and comparison pages go live for "what is it" and "how to choose". Acceptance: Q&A content indexed and appearing in the citation sources of relevant answers.
Month 3 · Mentions and re-measurement: work third-party mentions and directory profiles, re-measure on the same basis, set next quarter's priorities. Acceptance: sampling on a fixed question set and one method shows a trend, grounding next quarter's investment.
What a real diagnostic adds
- Raw measurements and evidence for your own domain, page by page — not a benchmark;
- A same-method comparison table against three competitors you name;
- A prioritised fix list (what first, why, expected impact);
- A walkthrough call covering the report item by item.
Fixed price RMB 19,800, no per-page or per-finding surcharges. Before ordering, run the free AI visibility audit for a rough picture; the diagnostic package page lists the full deliverables.
How to use this sample
Take this sample to any GEO provider and ask: does your deliverable look like this — scorecard with item-by-item evidence, findings with fixes, roadmap with acceptance criteria? A provider that cannot answer mostly delivers a monthly "work summary". The full vetting script: the 12 due-diligence questions.
Related questions: What does an AI visibility audit actually check? A complete checklist; How much does GEO cost? How is it usually priced?; What is GEO (Generative Engine Optimization)?