From Theory to Practice – LLMO and the AIDiscoverability Index (AIDI) – Part 2

Last updated: June 13, 2026

In the first part of this series, we explored the fundamental shift toward AI discoverability and defined the key terms LLMO, AEO, and GEO. But how do you measure success in this new world? How can you objectively assess how well your company is understood and recommended by artificial intelligence? The answer lies in a new benchmark: the AI Discoverability Index (AIDI), and an understanding of how modern AI systems actually work.

RAG Systems: The Bridge Between LLMs and the Real World

A common misconception is that language models like ChatGPT directly search the internet. Classic LLMs are limited to their static training data. However, modern systems like ChatGPT with a browsing feature or Google AI Overviews are more advanced. They are so-called Retrieval-Augmented Generation (RAG) systems [3].

A RAG system works in two steps:

Retrieval: When a query is submitted, an upstream search infrastructure searches external, up-to-date sources (e.g., search indexes, databases, your website) to find relevant information.

Augmented Generation: The language model receives this retrieved information as additional context and uses it to formulate a well-founded, up-to-date, and contextually appropriate response.

Infrastructure: Can the AI analyze your website and data technically correctly? This involves the machine-readability of your content, clean code, and structured data.

Perception: Do external signals confirm your brand’s reputation and expertise? The AI evaluates whether you are cited in industry media, how you are discussed in forums, and what sentiment is associated with your brand.

Commerce: Will the AI recommend your products with conviction? This pillar evaluates the quality of your product information, the availability of reviews, and the clarity of your unique selling points.

This hybrid approach is crucial because it reduces typical LLM weaknesses such as outdated knowledge or “hallucinations” and increases trustworthiness, as the answers are based on verifiable sources[3]. For your LLMO strategy, this means: Your content must be optimized for the retrieval step in order to even be shortlisted for generation.

AIDI: The Benchmark for Your LLMO Strategy

This is exactly where the AI Discoverability Index (AIDI) comes in. It measures not only whether you are mentioned, but also how well your content is prepared for this entire process. AIDI is a multi-agent framework that measures “Reasoning Readiness”—that is, a brand’s readiness to be understood and logically processed by AI—across 13 dimensions[2].

The Three Pillars of AIDI

For executives, this complex framework can be summarized in three simple pillars:

The 13 Dimensions: The Technical Depth of AIDI

Behind these pillars lies the true strength of AIDI: its technical depth. The framework analyzes 13 specific dimensions to paint a holistic picture of AI readiness.

These include, among others:

Schema & Structured Data: The use of standardized formats (e.g., from Schema.org) to explicitly define the meaning of content.

Semantics & Entities: The use of clearly defined terms and linking to known entities in knowledge graphs.

Sentiment Analysis: The analysis of the tone of external mentions and customer reviews.

Conversational Copy: The presentation of text in a natural, conversational style.

Knowledge Graph Presence: Anchoring your brand in public and private knowledge databases.

SentimentScore:
The average sentiment of mentions of your brand.

Recommendation Rate: The percentage of relevant queries in which your products are recommended.

Knowledge Panel Accuracy:
The accuracy and completeness of the information that appears in AI-generated summaries about your company.

This technical precision is AIDI’s “moat” and sets it apart from superficial keyword trackers. It’s not about whether a word appears, but whether the AI understands the concept behind it.

FAQ for Small and Medium-Sized Businesses (Part 2): Measurability and KPIs

1. How can we measure the AIDI for our company?

Measuring the AIDI requires specialized tools and a deep understanding of the framework. This is an area where external consultants play a crucial role. They can conduct an AIDI audit that reveals your company’s current status, compares it to the competition, and provides a clear roadmap for improvements. Such an audit is not a one-time snapshot, but rather the starting point for continuous optimization.

2. What new KPIs should we track?

In addition to the metrics already mentioned in Part 1, such as “citation rate” and “share of AI voice,” you should include qualitative KPIs in your reporting:

3. How quickly can we see results?

Improving your AIDI score is a marathon, not a sprint. While some technical optimizations (e.g., schema implementation) can lead to quick improvements, building authority (perception) is a long-term process. However, initial beta tests show that brands in the top AIDI quartile see over 40% more AI-driven traffic than the average [2]. So the investment pays off.

4. Why Conventional “AI Visibility Trackers” Are Not Enough

Many of the tools currently available focus on counting how often a brand is mentioned in AI responses. This is an important data point, but it is superficial. AIDI goes deeper and asks: Is the brand understood correctly? Is it mentioned in the right context? Is the mention positive or neutral? Is it perceived as a trustworthy source for specific topics? These qualitative dimensions are crucial for long-term success and can only be captured through a comprehensive framework like AIDI.

Preview of Part 3

In the third and final part of this series, we’ll show you how to design your LLMO roadmap using the PixelMechanics Total Experience approach and what specific on-page and off-page measures you can take.

References

[1] HubSpot. (2025). Best practices for answer engine optimization (AEO) that marketing teams can’t ignore. https://blog.hubspot.com/marketing/answer-engine-optimization-best-practices
[2] Parr, Dale. (2025). AIDI: The New Standard for AI Discoverability. Taken from the provided document.
[3] eology GmbH. (2025). Large Language Model Optimization (LLMO). https://www.eology.de/magazine/large-language-model-optimization

Frequently Asked Questions About the AI Discoverability Index (AIDI)

What is the AI Discoverability Index (AIDI)?

The AI Discoverability Index (AIDI) is a metric that measures how visible and citable a brand or website is in AI-powered search systems such as ChatGPT, Perplexity, and Google AI Overviews. Background information in the AI overview.

How do I measure my brand’s AI visibility?

By testing relevant prompts in various AI systems and evaluating how frequently, accurately, and prominently your brand is mentioned and linked—and tracking this value over time as an index.

How do I improve the AIDI with LLMO?

Through structured, authoritative content with clear answers, consistent entity information, reliable sources, and technical discoverability, so that AI systems reliably cite the brand as a source. Our AI consulting supports you in this process.