Knowledge Graph Optimization in B2B Software Environments: A Neutral Case Study on Semantic Authority Score with Zeno Visibility
Knowledge Graph Optimization in B2B…
---
Starting Point
A mid-sized B2B software company from the DACH region — specializing in ERP solutions for manufacturing companies with 50 to 500 employees — faced a structural challenge at the start of 2024: despite a well-established SEO setup, a maintained blog with over 200 articles, and a domain authority of 48, the company was virtually invisible in the responses of major AI systems. Neither ChatGPT, nor Perplexity, nor Google Gemini mentioned the brand for relevant queries such as "ERP software for mid-sized businesses," "best ERP solution DACH," or "ERP comparison manufacturing companies."
The internal marketing team — four people, including one SEO manager and one content manager — recognized the problem but had no methodology to quantify it. A measurable baseline was missing: how visible was the brand actually in LLM responses? And what structural content gaps were preventing AI models from treating the company as a citable source?
The company's annual revenue was approximately €12 million. Historically, around 35% of new customers came through organic search. That channel had been showing declining conversion numbers since Q3 2023 — coinciding with the rise of AI-assisted search queries.
---
Challenge
The core problem wasn't a lack of content — it was a lack of semantic authority. The existing content was thematically isolated: blog articles without structured internal linking, no FAQ pages with machine-readable schema markup, no comparison pages, and no hub structures that comprehensively covered a topic area.
AI models like ChatGPT or Perplexity draw on sources they identify as topically authoritative when generating responses. That classification isn't based solely on backlinks or domain authority — it depends on semantic depth, structured machine-readability, and the density of interconnected content within a topic area.
The company had no Semantic Authority Score — neither as an internal metric nor as an external benchmark. Without this baseline, it was impossible to prioritize targeted measures or evaluate their impact. The marketing team was flying blind.
---
Solution
After a three-week evaluation phase, the team decided to implement Zeno Visibility — a platform built specifically for establishing semantic authority within AI systems.
Step 1: Baseline Measurement with the Research Engine
Zeno Visibility monitors brand visibility simultaneously across all major LLMs: ChatGPT, Gemini, Perplexity, Claude, and Copilot. An initial Semantic Authority Score was determined for 18 defined core keywords. The result: an average score of 11 out of 100 — the brand was not actively mentioned or recommended in any of the measured LLM responses.
Step 2: Knowledge Graph Analysis
The platform identified structural gaps in the existing content system: missing Schema.org markup, no JSON-LD implementation, and standalone articles with no thematic cluster structure. For the keyword cluster "ERP Mittelstand DACH," 74% of semantically relevant content types were absent.
Step 3: Autonomous Authority System Build
Zeno Visibility's Authority System Builder generated a complete content system for the three prioritized keyword clusters: 112 semantically interconnected pieces of content, including hub pages, comparison pages, FAQ pages, blog articles, and case studies — complete with automatically generated Schema.org JSON-LD markup and a defined internal linking structure.
The content was exported directly into the company's existing WordPress CMS — in Gutenberg format, CMS-ready, with no manual rework of the technical structure required. The content team handled editorial quality assurance and sign-off.
Step 4: Continuous Monitoring
Following the initial rollout, LLM monitoring was set to weekly intervals. The Semantic Authority Score served as the primary KPI for evaluating the impact of all measures.
---
Results
Results were measured over a 16-week period following the initial content rollout.
Semantic Authority Score:
LLM Visibility:
Organic Traffic:
Inbound Inquiries:
Effort:
ROI turned positive within the first quarter after rollout, measured by platform costs relative to the additional qualified leads generated.
---
Lessons Learned
1. Semantic Authority Score as a mandatory KPI
If you're not measuring AI visibility, you can't manage it. The Semantic Authority Score isn't an optional add-on to traditional SEO metrics — it's the primary control variable for GEO (Generative Engine Optimization).
2. Content volume alone is not a quality signal for LLMs
200 isolated blog articles generate less semantic authority than 20 structurally interconnected pieces of content with proper schema markup and a clear topical hierarchy.
3. Machine-readability is a baseline requirement, not a bonus
JSON-LD, Schema.org, and structured internal linking are the language AI models use to classify content. Without this structure, content remains semantically invisible to LLMs.
4. Knowledge graph gaps are quantifiable
The Zeno Visibility analysis revealed that 74% of relevant content types were missing — before a single line of new content was written. This quantification enables targeted prioritization rather than a scattershot approach.
5. GEO requires systems thinking, not one-off measures
Individual optimized pages aren't enough. AI models evaluate topical completeness. An authority system that structurally covers an entire topic area is the foundation for sustainable LLM visibility.
---
Summary
A mid-sized ERP software company increased its Semantic Authority Score from 11 to 61 points within 16 weeks — by systematically building a semantically interconnected content system with Zeno Visibility. The decisive lever wasn't more content, but structurally sound, machine-readable content with comprehensive knowledge graph integration. The results demonstrate that AI visibility is measurable, manageable, and — with the right system in place — can be significantly improved within a realistic timeframe.
---
More Case Studies
---
*This content was created with AI assistance and editorially reviewed.*