AIVA Index
Less than 30 out of 100
Case Study
How a Swiss Pension Fund Doubled the Visibility of Its Google AI Overviews
By establishing an AI Visibility Architecture (AIVA)
Executive Summary
One of Switzerland’s leading pension funds was facing a problem that hardly anyone in the financial industry is aware of anymore:
Despite having an established brand, a functional website, and solid content, the company was virtually invisible in AI systems.
HR managers, CEOs, and policyholders are increasingly using AI systems to ask questions about retirement planning. Those who aren’t listed there lose credibility—even before the first conversation takes place.
+97 % Google AI Overview: Visibility
0 % → ~100 % LLM Visibility in Key Prompts
Level 3 AIVA Maturity Model (Levels 1–2)
6–8 weeks Implementation Period (AIVA Sprint)
Measured using Peec AI—across models, stable, and reproducible.
The Initial Situation:
The “Ghost Entity” in the Pension Segment
The pension fund had:
- An established brand with a long-standing market presence
- A well-organized website with relevant content
- An effective digital presence in the traditional sense
And yet:
- For ChatGPT, Perplexity, and Claude: hardly any mentions
- Google AI Overview: Visibility in Only 27.3% of Relevant Prompts
- For individual key prompts: 0% visibility—even when the brand name was included in the prompt
The analysis showed:
Although the website provided information, it did not contain any content that LLMs could use as a citation source.
In the AIVA model, this means:
No architecture → no visibility
The source data (Peec AI Analysis)
AIVA Maturity Level
Level 1–2 (Invisible / Fragmented Present)
Google AI Overview: Visibility
27.3% (for customer-defined prompts)
LLM Visibility (Key Prompts)
0% – not mentioned; no reference to the brand as the source
Core of the problem
No structured, citable content on the website
The real cause:
An architectural problem, not a content problem
The diagnosis based on the AIVA framework revealed three structural deficits
Lack of Responsiveness in the Content
LLMs don’t search for websites—they search for answers.
The pension fund had content, but none that the models could use directly to answer specific user questions. Neither structurally nor semantically were the texts organized in a way that allowed LLMs to recognize them as a citable source of facts.
AI systems were unable to provide clear, machine-readable answers on topics such as:
- BVG Fundamentals and Pension Solutions
- Differences Between Pension Plans
- specific benefit promises made by the pension fund
No External Validation Layer
AI systems do not rely primarily on company websites.
They rely on:
- Third-Party Sources and Reference Platforms
- Structured Knowledge Sources
- Externally Verified Information
→ The pension fund did not have sufficient external validation signals—even on platforms that classify LLMs as highly authoritative.
Missing Co-Occurrence Patterns
Competitors and institutions consistently appeared in connection with:
- Relevant technical terms (BVG, Pillar 2, portability)
- Specific Application Scenarios
- Institutional Contexts
→ These patterns were completely absent from the brand. For AI systems, this means there is a low probability of mentioning the brand—even in queries closely related to the brand.
The Solution: Implementing AIVA (AI Visibility Architecture)
Instead of scaling operational measures, the entire visibility strategy was reimagined:
Not as a content problem. But as an architectural problem.
The implementation took place as part of a focused 6–8-week sprint—not an ongoing agency retainer, but a one-time investment with a clearly defined goal: to understand why AI systems ignore the organization—and to build the architecture that changes that.
The implementation took place in four systematic steps:
Prompt Landscape Mapping
Step 1- Analysis of Relevant AI Demand in the Retirement and Pension Fund Segment
- Identification of Key Strategic Prompts (BVG Questions, Pension Options, Comparisons)
- Mapping the Decision-Making Questions of HR Managers, Executives, and Insured Individuals
- Prioritization Based on Relevance for Building Trust and Influencing Customer Decisions
AI Mention Mapping
Step 2- An analysis of which brands and institutions dominate AI responses—and why
- Identifying the Sources That LLMs Rely On for Health Prevention Topics
- Identifying Co-Occurrence Gaps: Where Was the Brand Missing in the Context of Relevant Technical Terms?
- Competitor positioning in AI systems is systematically tracked
Architecture Rebuild (Core AIVA Lever)
Step 3- Establishment of a structured knowledge center with content that provides answers
- Semantic Topic Clusters: Content structured so that LLMs recognize it as a direct source of answers
- Entity Authority: Clear Linkage Between Brand, Subject Areas, and Specific Pension Planning Use Cases
- Knowledge Graph Signals: Machine-Readable Fact Clusters for LLM Citation
- Result: The Knowledge Center is demonstrably used as a source (as measured by Peec AI)
Authority Amplification
Step 4- Strategic reference content that AI systems actively cite as a source
- External validation signals on relevant platforms in the healthcare sector
- Targeted long-tail coverage for questions relevant to decision-making
- Not through more content—but through more precise architectural cues
The Results: From Invisibility to Structured AI Visibility
Within 6–8 weeks after implementing the AIVA Architecture Sprint:
| Before | After | |
|---|---|---|
| Google AI Overview Visibility: | 27,3 % | 53,8 % |
| LLM Visibility in Key Prompts: | 0 % | ~100 % |
| AIVA Maturity: | Level 1–2 | Level 3 |
| No content that can be quoted or responded to | The Knowledge Center is demonstrably used as a source |
27,3 % → 53,8 %Google AI Overview: Visibility
0 % → ~100 %LLM Visibility in Key Prompts
Level 1–2 → 3AIVA Maturity
Level
Level
Source Authority EstablishedThe Knowledge Center is used as a source of citations
Why This Strategy Works
AI systems work fundamentally differently from search engines.
They are based on
- Reference Structures and Semantic Relationships
- Externally validated sources of knowledge
- Content that generates responses—not just pages that rank well
Companies that are actively shaping this architecture
- Are mentioned more frequently on a systematic basis
- Are classified as a trusted source
- We’re there before the first meeting with a client takes place
The Strategic Impact
The biggest impact isn’t traffic.
Rather: AI Authority.
In the context of retirement planning, this means
- Greater likelihood of being mentioned during purchasing processes and when switching providers
- A stronger reputation as a reputable, competent provider
- Building Trust Even Before the First Contact
- A structural competitive advantage as long as competitors continue to rely on traditional SEO
Particularly Relevant in the Pension Fund Segment
- HR managers and CEOs are increasingly using AI systems to research retirement planning solutions. If a company isn’t listed there, it simply doesn’t exist for the new generation of decision-makers.
Our Conclusion
This project serves as a prime example of something that many companies in the DACH region have not yet grasped:
AI Visibility ist kein Content- oder SEO-Thema. Es ist ein Architektur-Thema.
Before investing in AI Visibility, the current situation must be clear—and it is important to understand why AI systems are ignoring an organization.
→ That’s exactly what the AIVA Architecture Sprint does.
If you start without this assessment, you’re optimizing for nothing.
Companies that are actively building their AIVA (AI Visibility Architecture) will:
- Mentioned more frequently
- Found Faster
- Als relevanter wahrgenommen
All the others will:
- → invisible
