Last updated: June 22, 2026
In the first part of this series we examined the fundamental shift toward AI visibility and placed the key terms LLMO, AEO and GEO in context. 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 in understanding how modern AI systems actually work.
RAG systems: the bridge between the LLM and the real world
A widespread misconception is that language models like ChatGPT directly search the internet. Classic LLMs are limited to their static training data. Modern systems such as ChatGPT with browsing capability or Google AI Overviews, however, are more advanced. They are so-called Retrieval-Augmented Generation systems (RAG) [3].
A RAG system works in two steps:
Retrieval: When a query is made, 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, on this basis, formulates a well-founded, up-to-date and context-appropriate answer.
- Infrastructure: Can the AI technically analyze your website and your data correctly? This is about the machine readability of your content, clean code and structured data.
- Perception: Do external signals confirm the reputation and expertise of your brand? The AI evaluates whether you are cited in trade media, how you are talked about 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This hybrid approach is crucial, because it reduces typical LLM weaknesses such as outdated knowledge or “hallucinations” and increases trustworthiness, since the answers are based on verifiable sources[3]. For your LLMO strategy, this means: your content must be optimized for the retrieval step to even make the shortlist 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 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 an AI – across 13 dimensions [2].
The three pillars of AIDI
For executives, the complex framework can be summarized in three simple pillars:
The 13 Dimensions: The Technical Depth of AIDI
Behind these pillars lies the real 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 make the meaning of content explicit.
- Semantics & Entities: The use of clearly defined terms and the linking to known entities in knowledge graphs.
- Sentiment Analysis: The analysis of the tonality of external mentions and customer reviews.
- Conversational Copy: The preparation of texts in a natural, dialogue-oriented style.
- Knowledge Graph Presence: The anchoring of one's own brand in public and private knowledge databases.
- Sentiment Score:
The average tonality of the mentions of your brand. - Recommendation Rate: The percentage share with which your products are recommended in relevant queries.
- Knowledge Panel Accuracy:
The correctness and completeness of the information that appears in AI-generated summaries about your company.
This technical precision is the "moat" of AIDI and distinguishes it from superficial keyword trackers. It is not about whether a word appears, but whether the AI understands the concept behind it.
FAQ for SMEs (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 in which external consultants play a decisive role. They can conduct an AIDI audit that shows the current status of your company, compares it with the competition and delivers a clear roadmap for improvements. Such an audit is not a one-time snapshot, but the starting point for continuous optimization.
2. Which 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 the AIDI score is a marathon, not a sprint. While some technical optimizations (e.g. schema implementation) can quickly lead to improvements, building authority (Perception) is a long-term process. Initial beta tests, however, show that brands in the top AIDI quartile record over 40% more AI-driven traffic than the average [2]. So the investment pays off.
4. Why are conventional "AI visibility trackers" not enough
Many of the currently available tools focus on counting how often a brand is mentioned in AI answers. That is an important data point, but it is superficial. AIDI goes deeper and asks: Is the brand correctly understood? 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 decisive for long-term success and can only be captured by a comprehensive framework like AIDI.
Outlook on Part 3
In the third and final part of this series we show you how to design your LLMO roadmap with the PixelMechanics Total Experience approach and which concrete on-page and off-page measures you can take.
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.
How do I measure my brand's AI visibility?
By testing relevant prompts across different AI systems and evaluating how often, accurately, and prominently your own 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, strong sourcing, and technical discoverability, so that AI systems reliably draw on the brand as a source.
What does the AI Discoverability Index measure?
How well a company is understood, described and recommended by AI systems. It combines whether you are found at all, how accurately you are described, and in what context you get named.
How is it measured in practice?
With a fixed set of prompts, asked repeatedly and evaluated the same way each time. The absolute number matters less than the direction of travel across quarters.
What influences the result most?
Clarity. Companies that describe what they do in plain language, in one place, get summarised correctly. Those that describe themselves in marketing abstractions get summarised as something else entirely.
