PageRank, Domain Authority, Core Web Vitals — crypto site owners have tracked these metrics for years. In 2026, there’s a new number that matters more for discovery: AVI Score, the AI Visibility Index.
AVI Score measures how visible your site is to AI systems — ChatGPT, Claude, Perplexity, Gemini, and 18 others — on a 0 to 10 scale. Unlike traditional SEO metrics, AVI Score predicts AI citation probability: the likelihood that when someone asks an AI assistant about your topic, your site gets mentioned.
Why AI Citation Is the New Organic Traffic
Consider the numbers. ChatGPT now processes 10M+ crypto-related queries per week. Perplexity handles 100M+ daily searches and cites sources inline. Google AI Mode — live in 2026 — generates direct answers from crawled content. Users who get an answer from these systems rarely click through to Google results at all.
Sites not cited by AI lose this traffic entirely — not because Google demoted them, but because AI systems never learned they existed.
How AVI Score Works
The score is calculated from 36 technical signals across five categories:
- Crawlability (10 signals): Are AI bots allowed in robots.txt? Does llms.txt exist? Is ClaudeBot, GPTBot, OAI-SearchBot, Google-Extended accessible?
- Schema (8 signals): Article JSON-LD, Speakable Schema, FAQ schema, NewsArticle type, author Person, Organization markup
- Local/Brand (7 signals): LocalBusiness JSON-LD, NAP consistency, brand entity signals, GMB data
- Technical (6 signals): HTTPS, page speed, Core Web Vitals, IndexNow, sitemap freshness
- Content (5 signals): Heading structure, content length, update frequency, internal linking
Each signal is checked automatically against the live site and scored pass/fail/warn. The composite AVI Score is weighted toward crawlability and schema signals, which have the highest correlation with actual AI citations.
btcnews.biz AVI Audit Results
We ran btcnews.biz through the full 36-signal audit. Initial score: 4.2/10. The audit surfaced three critical issues immediately:
- Missing llms.txt: No explicit AI permission file. ClaudeBot still crawled the site (43,000 requests in 20 hours — confirming AI interest), but without llms.txt, it had no structured guidance on what to cite.
- Absent Speakable Schema: None of the 110,000+ articles had SpeakableSpecification markup. This is the schema type specifically designed to signal citation-worthy text to voice AI systems.
- robots.txt gap: OAI-SearchBot (ChatGPT’s live web search bot, separate from GPTBot) wasn’t explicitly allowed.
After fixes: llms.txt deployed, Speakable Schema added to all articles via mu-plugin, robots.txt updated. Projected AVI Score: 6.8/10.
Checking Your Site’s AVI Score
The full 36-signal audit is available through AI Visibility Suite, a WordPress plugin that runs the complete check automatically and shows results in a Yoast-style dashboard. Starting at $49/year, it monitors 8 AI systems continuously and alerts you when your citation score drops.
Non-WordPress sites can use the standalone dashboard at api.visibility.webcorelab.com/v1/dashboard or embed the JS widget directly.
For a quick manual check: search your domain name in ChatGPT, Claude, and Perplexity. Count how many of 10 test queries return your site as a cited source. Below 3/10 means your AVI Score is likely below 5.0 and you’re leaving AI-native traffic on the table.
The Benchmark: What a Good AVI Score Looks Like
Based on audits of 200+ sites across verticals:
- 7.0–10.0: Consistently cited by top AI systems. Sites like major crypto exchanges and top-10 news sites.
- 5.0–6.9: Cited for brand queries, occasionally for topical queries. Typical for established crypto media with partial AI optimization.
- 3.0–4.9: Only cited for exact brand name searches. Most crypto news sites fall here.
- Below 3.0: Essentially invisible to AI. Often has AI bots blocked in robots.txt or zero schema markup.
The average site can move from the 3–4 range to 6–7 with one day of technical setup. The gap between AI-visible and AI-invisible sites in crypto content is almost entirely technical, not editorial.