# AI-visibility check — notion.so

**Score: 63/100 — Partial visibility** (checked with https://geoplaybooks.com/check/)

AI engines can reach you, but 3 signals need work to get you into most generated answers. Fix the WARN and FAIL items below.

| Signal | Status | Finding |
|---|---|---|
| AI-crawler reachability | PASS | Homepage returns HTTP 200 with ~2,686 chars of server-rendered text. Retrieval bots can read you. |
| GPTBot / ClaudeBot access | PASS | No blanket Disallow for the major AI crawlers — GPTBot, ClaudeBot and PerplexityBot can fetch your key pages. |
| llms.txt | WARN | /llms.txt is present but malformed — it should be markdown starting with an # H1 title and curated links. |
| JSON-LD schema | FAIL | No JSON-LD schema on the homepage. Engines can't tell what type of entity you are. |
| Answer-first structure | WARN | Headings exist but the copy under them is long-form. Add a 40–80 word answer capsule right after each heading. |

## AI fix-it prompt

```text
You are an engineer working on the website notion.so. Below is an AI-visibility (GEO) audit of the site — the signals that decide whether AI answer engines (ChatGPT, Perplexity, Google AI Overviews) can crawl, understand and cite it. Implement the fixes.

## Audit findings

### Sitewide signals
- [PASS] AI-crawler reachability: Homepage returns HTTP 200 with ~2,686 chars of server-rendered text. Retrieval bots can read you.
- [PASS] GPTBot / ClaudeBot access: No blanket Disallow for the major AI crawlers — GPTBot, ClaudeBot and PerplexityBot can fetch your key pages.
- [WARN] llms.txt: /llms.txt is present but malformed — it should be markdown starting with an # H1 title and curated links.
- [FAIL] JSON-LD schema: No JSON-LD schema on the homepage. Engines can't tell what type of entity you are.
- [WARN] Answer-first structure: Headings exist but the copy under them is long-form. Add a 40–80 word answer capsule right after each heading.

## Implementation rules
- Work through the failing signals in the order above; PASS items need no changes.
- Crawler access: check robots.txt AND the WAF/CDN layer for blocks on GPTBot, OAI-SearchBot, ClaudeBot, PerplexityBot, Google-Extended.
- llms.txt: create a curated Markdown index at /llms.txt per https://llmstxt.org — one H1, a > summary, ## sections, every link with a one-line description.
- JSON-LD: add Organization (+ logo, sameAs) sitewide and Article/FAQPage on content pages; link entities with absolute @id references.
- Answer-first structure: under each question-shaped H2, place a self-contained 40–80-word answer paragraph before any elaboration.
- Do not fabricate content; restructure what exists.

## Acceptance criteria
- Re-run the free checker at https://geoplaybooks.com/check/ for notion.so until all 5 signals are PASS.
- Validate llms.txt at https://geoplaybooks.com/tools/llms-txt-validator/ (0 errors).
- Verify JSON-LD parses in Google's Rich Results Test.
```

Deeper: page-by-page audit from $49, monthly monitoring from $20/mo — https://geoplaybooks.com/audit/