Two views below. Attributed wins = engines started citing us on a query AFTER we deployed a page targeting that exact query (proof of causality). Baseline citations = pre-existing brand mentions we did not cause.
Attributed Wins 0
A deployed page on the left, the AI engine's citation on the right. This is the closed loop.
Baseline Citations 0
Existing brand mentions across AI engines. These are not "wins" we caused, but they show engines already recognize this brand.
Evidence Trail
How AI Engines Found Us
Each card below is a verified instance where an AI engine cited this brand. Click any card to expand the raw AI response.
Before → After Evidence
Queries Where Citations Changed
Citation Scan
Every Query. Every Engine.
Are AI engines citing your content? Click below to replay the latest scan or run a fresh one.
citation scan
Ready. Click a button above to start.
0
Checks Run
0
Citations Found
0
Engines
0
Queries
Live Verification
Run a Query Right Now
Don't take our word for it. Pick a query and an AI engine, then watch it respond — live, unscripted.
live api call
Progress Timeline
How Citation Rate Moved Over Time
Remediation Pages
Pages AIsubtext Deployed 0
Each page is crafted to answer a specific query and get picked up by AI engines.
Recommended Actions
What You Can Do to Boost Visibility
These are channels where your brand is missing or underrepresented. Each card includes ready-to-use content you can publish with minimal effort.
How It Works
3-Layer AI Engine Fingerprinting
Most analytics tools rely on referrer headers, which AI engines often don't send. AIsubtext uses a novel 3-layer detection cascade to identify AI-origin traffic even when traditional signals are missing.
Layer 1
Referrer Headers
When an AI engine links to our content, the HTTP referrer reveals the source. We match against known AI engine domains.
Confidence: 95%
Layer 2
User-Agent Bot Signatures
AI engines use crawlers with distinctive User-Agent strings. We identify GPTBot, PerplexityBot, ClaudeBot, and others — even when referrers are empty.
Confidence: 90%
Layer 3
Behavioral Fingerprinting
When both referrer and UA are inconclusive, we score 7 behavioral signals: missing Accept-Language, no cookies, no Sec-Fetch headers, and more.
Confidence: 45–80%
Why this matters: Google Analytics shows you where web traffic comes from. AIsubtext shows you where AI recommendation traffic comes from — a metric that doesn't exist anywhere else.
Self-Improving System
Closed-Loop Feedback — Content That Learns
AIsubtext doesn't just deploy content and hope. Each remediation cycle feeds performance data back into the next generation of content. The system gets smarter with every scan.
1
Scan
Measure visibility across engines
→
2
Generate
Create counter-move pages
→
3
Track
Monitor AI referrals & citations
→
4
Learn
Feed results into next cycle
↺
What the feedback loop does:
• Replicates what works: Pages that got cited by AI engines inform the style/structure of future content
• Drops what doesn't: Pages deployed 14+ days ago with zero citations are flagged for revision
• Skips won queries: Queries already highly cited are deprioritized to focus effort on gaps
• Adapts per engine: Performance data per engine tunes content for each AI model's preferences
This is the moat. Anyone can deploy content. Only AIsubtext closes the loop — scan, remediate, track, learn, repeat.
This evidence is generated automatically by AIsubtext.