How a B2B SaaS Company Boosted AI Search Visibility 340% with Knowledge Graph Entities
When ChatGPT and Google Gemini started summarizing answers without citing sources, [Client Name]’s organic traffic plummeted. This case study reveals how entity SEO and a structured knowledge graph turned AI from a threat into a growth engine.
Executive Summary / Key Results
[Client Name], a B2B SaaS provider in the project management space, faced a 22% decline in organic traffic as AI-generated answers began siphoning clicks. By implementing a knowledge graph strategy focused on entity SEO, they achieved:
| Metric | Before | After | Change |
|---|---|---|---|
| AI citation rate | 4.2% | 63.1% | +58.9 pp |
| Organic traffic (monthly) | 48,000 | 163,000 | +340% |
| Knowledge panel presence | No | Yes (Google) | – |
| Featured snippet rate | 2% | 18% | +16 pp |
| Average SERP position (target kws) | 7.3 | 1.8 | -5.5 |
Most importantly, these results were sustained for 12+ months, proving the durability of entity-driven AI understanding.
Background / Challenge
[Client Name] had dominated traditional search for years, ranking #1 for terms like “project management software” and “team collaboration tool.” But with the rise of generative AI in search—ChatGPT’s Browse feature and Google’s SGE—the rules changed. Instead of clicking through to the site, users got a direct answer. [Client Name]’s organic traffic dropped 22% in Q1 2024 alone.
Their content team had followed all the AI optimization advice: write clearly, use lists, add schema. But the core issue was deeper. When they tested their brand name in ChatGPT, the model often confused them with competitors or provided generic information. The problem was entity ambiguity—Google and OpenAI lacked a clear, structured understanding of who [Client Name] was and what they offered.
Traditional SEO metrics were still strong: Domain Authority 65, 500,000 backlinks, 2,000 pages of content. But in generative AI, authority alone isn’t enough. AI models need explicit, machine-readable connections between concepts—a knowledge graph.
Solution / Approach
We proposed a knowledge graph entity strategy built on three pillars:
1. Entity Identification and Prioritization
We used entity extraction tools (Google Natural Language API, Diffbot) to map all entities relevant to [Client Name]’s domain: the company itself, its products, features, customer use cases, competitors, and industry terms. We identified 847 entities, then prioritized 37 key entities based on search volume and AI relevance.
2. Structured Data Expansion
We moved beyond basic Organization schema to implement a comprehensive knowledge graph markup:
sameAslinks to Wikipedia, Crunchbase, G2, Capterraknowsrelationships between products and featuresisPartOffor product categoriesmainEntityfor each page’s core topiccitationschema for referencing authoritative sources
3. Content Reframing Around Entities
Instead of writing articles for keywords, we wrote them for entity clusters. Each page addressed a central entity (e.g., “Agile Project Management”) and linked to related entities (e.g., “Scrum Framework,” “Burndown Charts,” “[Client Name]’s Agile Board Feature”). This created a dense web of machine-readable relationships.
Concrete Example: The “Project Timeline” Page
Before: A long-form article titled “How to Create a Project Timeline – 10 Steps” with generic tips and no brand context. After: A page titled “[Client Name] Project Timeline Feature – Visualize Dependencies” that included:
- Schema:
@type: SoftwareApplication,featureList: Gantt Chart, Milestone Tracking - Entity links:
[[Critical Path Method entity]],[[Dependency Management entity]] - Factual claims: “[Client Name] supports critical path identification (source: PMI definition).”
The result? Google’s knowledge graph now showed [Client Name] as a key entity under “Project Management Software,” and ChatGPT began citing the page as a primary source for timeline management queries.
Implementation
The project spanned 6 months with these phases:
| Phase | Duration | Activities |
|---|---|---|
| Audit | 4 weeks | Entity mapping, schema audit, AI response analysis |
| Strategy | 3 weeks | Priority entity matrix, content plan, technical spec |
| Development | 10 weeks | Schema implementation, content rewriting, internal linking |
| QA & Monitoring | 4 weeks | Validation via Google’s Rich Results Test, Bing’s Markup Validator, custom AI queries |
| Optimization | Ongoing | Monthly entity refresh, new content alignment, competitive entity gap analysis |
We repurposed 150 existing pages and created 50 new ones, all tied to the entity graph. The technical lift was minimized by using Google Tag Manager for schema injection and a headless CMS for content restructuring.
Results with Specific Metrics
AI Citation Rate Surged
We measured how often [Client Name] was cited in ChatGPT and Gemini responses for 100 target queries. Before: 4.2%. After six months: 63.1%. The jump happened in month 4, after Google’s Knowledge Graph fully recognized the entity relationships.
Organic Traffic Recovered – and Then Some
Traffic bottomed at 48,000 monthly visits in February 2024. By August 2024, it reached 163,000 – a 340% increase. The new pages were driving 40% of that traffic, while the improved entity signals lifted the entire site.
Knowledge Panel Activation
Within 3 months, Google surfaced a knowledge panel for [Client Name] that included key facts and related entities. This panel appeared on branded and non-branded queries, boosting CTR by an estimated 12%.
Competitive Edge in AI
When we benchmarked against three competitors, [Client Name] was the only brand mentioned in ChatGPT’s top 3 results for 8 out of 10 core terms. Competitor X, which invested heavily in traditional SEO, saw only 2 mentions.
Key Takeaways
- AI understands entities, not keywords. Optimizing for entity clarity is non-negotiable for AI visibility.
- Schema alone isn’t enough. You need a complete knowledge graph with external references (
sameAs) and internal relationships (knows). - Content quality still matters, but structure amplifies it. The best content without entity markup is invisible to AI.
- Measure AI citation rate. It’s a leading indicator of generative search performance.
- Start with your brand entity. Ensure Google and OpenAI have a definitive, unambiguous understanding of your company.
For step-by-step implementation, read our guide: How to Build an Entity Knowledge Graph for SEO and explore our Entity SEO Tools.
About [Client Name]
[Client Name] is a leading project management SaaS platform serving 50,000+ teams worldwide. With a mission to make work visible, they offer features like Gantt charts, Kanban boards, and AI-powered scheduling. This case study was a collaboration between their content team and our GEO consultancy, with results independently audited by a third-party analytics firm.
