By William Wang, Founder of GEOScore AI
This is the story of how a mid-sized B2B SaaS company went from being completely invisible to AI search engines to being regularly cited by ChatGPT, Perplexity, and Google's AI Overviews — in 8 weeks.
The company (which asked to remain anonymous) sells project management software. They had strong traditional SEO — ranking on page 1 for several competitive keywords. But when they tested queries about their product category in ChatGPT and Perplexity, they were never mentioned. Their competitors were.
The Starting Point: GEO Score 23/100
When they ran their first scan through GEOScore AI, the results were sobering:
- Overall GEO Score: 23/100 (Grade: F)
- AI Crawler Access: Blocked — robots.txt was blocking GPTBot, ClaudeBot, and PerplexityBot
- llms.txt: Missing entirely
- Schema Markup: Only basic Organization schema, no Article or FAQ markup
- Content Structure: Long-form content buried answers under lengthy introductions
- Author Signals: No author bylines on blog posts
The diagnosis was clear: their site was technically invisible to AI search engines, and even if crawlers could access it, the content was not structured for citation.
Week 1-2: Technical Foundation
Fix 1: robots.txt (Impact: Critical)
They updated robots.txt to explicitly allow all major AI crawlers. This single change was the most important — without it, nothing else would matter.
Before:
User-agent: *
Disallow: /api/
Disallow: /admin/
# No AI-specific rules = default allow, BUT...
# Their Cloudflare WAF was blocking unknown user agents
After:
User-agent: GPTBot
Allow: /
User-agent: ChatGPT-User
Allow: /
User-agent: ClaudeBot
Allow: /
User-agent: PerplexityBot
Allow: /
User-agent: Google-Extended
Allow: /
User-agent: *
Disallow: /api/
Disallow: /admin/
They also updated their Cloudflare WAF rules to whitelist AI crawler user agents.
GEO Score after Week 2: 41/100
Fix 2: llms.txt (Impact: High)
They created a comprehensive llms.txt file listing their product pages, key blog posts, documentation, and case studies.
Time to implement: 30 minutes
Week 3-4: Content Restructuring
Fix 3: Direct Answer Format
They restructured their top 15 blog posts to lead with direct answers. Instead of:
"In today's fast-paced business environment, project management has become increasingly complex..."
They rewrote openings like:
"The best project management methodology for remote teams in 2026 is a hybrid approach combining async Kanban boards with weekly synchronous sprint reviews. Here is why, and how to implement it."
Fix 4: FAQ Sections
They added FAQ sections with Schema.org FAQPage markup to their 10 most-visited pages. Each FAQ contained 5-8 questions that mapped to real queries users ask AI search engines.
Fix 5: Factual Density
They audited content for vague claims and replaced them with specifics:
- "Improves productivity" → "Reduces project delivery time by 31% based on data from 2,400 teams"
- "Many companies use our tool" → "Used by 8,500+ teams including 12 Fortune 500 companies"
GEO Score after Week 4: 62/100
Week 5-6: Authority Building
Fix 6: Author Bylines
Every blog post got a proper author byline with credentials, photo, and link to the author's LinkedIn. They also created individual author pages.
Fix 7: Schema Markup Expansion
They added:
- Article schema to all blog posts (with author, datePublished, dateModified)
- FAQPage schema to FAQ sections
- Organization schema with comprehensive company details
- Product schema to product pages
GEO Score after Week 6: 74/100
Week 7-8: Optimization and Monitoring
Fix 8: Internal Linking
They improved internal linking to ensure their most important content was discoverable within 2 clicks from the homepage.
Fix 9: Content Freshness
They updated dates on all evergreen content and added "Last updated" timestamps.
Final GEO Score: 87/100 (Grade: A)
The Results
Within 4 weeks of reaching GEO Score 87:
- ChatGPT started citing their blog posts in responses about project management best practices
- Perplexity began including their product in comparison queries ("best project management tools for remote teams")
- Google AI Overviews started featuring snippets from their FAQ sections
They estimated the AI search traffic represented approximately 15% additional organic traffic on top of their existing traditional search traffic — and this percentage was growing monthly.
Key Takeaways
robots.txt is the #1 blocker — If AI crawlers cannot access your site, nothing else matters. This accounted for the biggest single score jump (23 → 41).
Content structure matters more than content quality — Their content was already high quality. The problem was that it was not structured for AI extraction.
Small changes compound — No single fix transformed their score. It was the combination of 9 targeted improvements that moved them from 23 to 87.
Speed matters — The sooner you fix technical blockers, the sooner AI systems start indexing your content. They wished they had started 6 months earlier.
Measurement drives improvement — Running regular GEOScore scans after each change let them see which fixes had the most impact and prioritize accordingly.
Your Turn
Want to know your GEO score? Run a free scan at geoscoreai.com — takes 60 seconds, no signup. You will get your score, top 3 blocking issues, and a clear starting point for improvement.
William Wang is the founder of GEOScore AI. He writes about AI search optimization and helping websites get found by the next generation of search engines.
Top comments (0)