What an AI visibility report looks like
An anonymised sample: the website of an automotive parts plant, 13 checks, total score 61 (grade C). Company name, domain and exact figures are anonymised or given as ranges; this is not a promise of results elsewhere.
The report has four parts
1. Total score and grade
Out of 100: A ≥ 85, B ≥ 70, C ≥ 50, D < 50. The score measures whether a machine can read and understand the page — not design quality.
2. Four category scores
Readability 30, third-party evidence 25, content coverage 20, AI visibility 25. You can see at a glance which one drags.
3. All 13 checks
Each with the evidence and measured value: body text length, number of JSON-LD blocks, canonical target.
4. Three priority fixes
Ordered by cost, capped at three. Fix, then retest — too many changes at once hides cause and effect.
Report body (anonymised)
Excerpts from the report. Figures are given as ranges and the client is not named.
更新日期 2026-11-05
1. Sample profile
Industry: automotive parts (precision machining + assembly) | Location: a prefecture-level city in East China | Size: two plants, about 400 staff.
Scanned: homepage and three product pages | Method: server-side fetch of public pages, no login, no form submission | Date: October 2026.
Note: this sample illustrates report structure only. Client name, domain and line-level data are anonymised.
2. Total score and four categories
Total 61 / 100, grade C (clear gaps).
Site readability 22 / 30: homepage returns HTTP 200 over HTTPS, but readable body text is only about 380 characters; specifications live in images and PDFs that crawlers cannot read.
Third-party evidence 12 / 25: no JSON-LD, no Organization or Product entity; og:title present, og:description missing.
Content coverage 15 / 20: title 32 characters (fine), description 18 characters (short), one H1 but only 4 characters.
AI visibility 12 / 25: sitemap has 6 entries and is not declared in robots.txt; canonical points to the old non-www domain; viewport present.
3. Three priority fixes
1. Move specifications from images into body text: target 800+ characters on the homepage and 600+ per product page. About 3 working days.
2. Add Organization and Product JSON-LD: legal name, address, phone, product models and scope. About 1 working day.
3. Normalise canonical to the www domain, declare the sitemap in robots.txt, and extend it to every active product page. About 0.5 working day.
4. Retest and limits
Retest the same 13 checks after 30 days and compare each item's status, not just the total. In our experience the first three fixes typically move the total by 15–25 points, but that is not a commitment.
This report does not include live testing in individual AI platforms and does not predict indexing, ranking or recommendation probability. Platform testing is scoped separately.
The report and raw scan data are used for this conversation only, are not published, and contain no identifiable client information.
5. Get your own report
Self-service: open /ai-visibility, enter your domain, and get the same 13 checks in seconds.
With a human: email the domain and your question to tony_2003@163.com, or add WeCom MES-3D (note "AI diagnosis"). We sequence the fixes within 30 days by cost.
PDF sample: Download the AI Visibility Diagnosis Sample Report (PDF) to forward to colleagues or use for internal review.