From 0% to 45%How Sighted Improved Its Own AI Visibility in 120 Days
120 days of self-experimentation. Interventions mapped to Citation Rate outcomes in a full timeline.
0%
Starting Citation Rate
45%+
Citation Rate at Day 120
120
Days
22+
Articles published
Why We're Publishing Our Own Case Study
Improving AI Visibility is uncharted territory for most companies. The question "what do you actually do to get cited by AI?" needs practical data, not just theory.
Sighted used its own domain (lp.sighted-aeo.com) as a test bed, practicing AI Visibility improvement over 120 days. This case study documents the entire journey as a causal chain: intervention, measurement, result.
Individual intervention details are in the AEO Optimization Log series (#001-#007). This article synthesizes them into a single 120-day narrative.
120-Day Timeline
Baseline measurement
Set up 109 questions. Brand query Mention Rate: 100%. Industry query Citation Rate: 0%.
Citation Rate 0%
Technical foundation
Implemented structured data (Organization, WebSite, Article). Optimized robots.txt. Added llms.txt. Improved site speed.
Citation Rate 0% — no change
Content expansion Phase 1
Published 8 new knowledge articles targeting industry queries. Self-contained structure, 2,000-4,000 words each.
Citation Rate 0% → 3%
Entity strengthening
Major About page overhaul. Author profile enrichment. Schema.org Person-Organization linking. Explicit ontology design.
Citation Rate 3% → 12%
Content expansion Phase 2
Published 4 industry guides (SaaS, EC, Local, Publisher). Case study and experiment data articles. Added 20 glossary terms.
Citation Rate 12% → 28%
Question Space coverage expansion
Created content for gap questions discovered via GPR experiment. External media contributions. Bilingual blog posts.
Citation Rate 28% → 45%
Continuous optimization
Weekly monitoring and intervention validation. New question discovery and response. Benchmark data accumulation.
Citation Rate 45%+ (ongoing)
What Worked and What Didn't
High Impact
Self-contained content
Each article fully answers a question. AI cites one page instead of stitching from multiple sources.
Entity clarification
Schema.org Organization + Person implementation. Made "What is Sighted?" accurately answerable by AI.
Question Space gap-filling
Prioritized content for questions where GPR experiment showed 0% citation.
Publishing proprietary data
109-question experiment results, scoring algorithm details -- unique data that no competitor has.
Limited Impact
Technical SEO alone
Structured data and site speed are necessary but insufficient. Citation Rate did not move from tech fixes alone.
llms.txt deployment
Useful as a crawler signal but no measurable Citation Rate improvement from llms.txt alone.
Short article volume
Articles under 500 words rarely get cited by AI. Volume at the expense of depth is counterproductive.
Backlinks (short-term)
Backlink effect was not as immediate as in SEO. AI trust-building takes longer.
Five Lessons from 120 Days
1. Content depth is the primary driver
Over 70% of Citation Rate improvement came from content quality and depth. Technical improvements are necessary but not sufficient.
2. Expect a 2-4 week lag
After publishing content, it takes 2-4 weeks for AI responses to reflect it. Do not judge by a single measurement; observe continuously.
3. Maturity stages cannot be skipped
Building the foundation at each maturity stage before advancing to the next is ultimately faster than trying to jump ahead.
4. Question Space visualization drives strategy
Mapping the Question Space quantifies gaps. This shifts from intuitive to data-driven prioritization.
5. Proprietary data is the strongest differentiator
Publishing unique data (GPR experiment, scoring algorithm) accelerated Citation Rate. AI preferentially cites information available nowhere else.
Start Improving Your AI Visibility
Begin with a baseline diagnosis. Enter your domain to check your Citation Rate.
Try free diagnosis