Frequently Asked Questions

AI SEO Tool by KE measures how visible brands are inside Large Language Model responses, using three core metrics: Latent Brand Association (the model's built-in associations with a brand), LLM Authority Score (how frequently and prominently a brand appears in AI responses), and Top of Mind (whether the model recalls a brand without external retrieval). The directory scores on this site were measured against OpenAI's GPT models; the metrics themselves apply to any LLM.

We queried the model many times with relevant prompts for each industry and brand, in both "recall" mode (no web access, testing what the model inherently knows) and "retrieval" mode (with live web search), and then analyzed the responses for brand mentions, positioning, sentiment, and associations. Each metric is scored on a scale of 0 to 100, where higher scores indicate stronger AI visibility. The multi-query approach accounts for the probabilistic nature of AI outputs, giving a reliable score rather than a single-point snapshot.

Each model has different training data, training cutoff dates, and retrieval systems, so a brand's visibility can vary significantly from one LLM to another. For example, a brand might score well on ChatGPT but poorly on Gemini because of differences in what content each model was trained on, how recently its data was updated, and how its retrieval-augmented generation (RAG) system works. The scores in this directory reflect one model family (OpenAI's GPT), so treat them as one model's view of the brand landscape. The principles behind the metrics apply to every LLM.

The Overall AI Visibility Score is a composite of our three core metrics (Latent Brand Association, LLM Authority Score, and Top of Mind), giving you a single number from 0 to 100 that represents your brand's total AI visibility. It's designed to be the one number you track over time. The formula is a smoothed geometric mean with an LBA-based floor on Authority and TOM, so a brand the model clearly recognises (high LBA) but doesn't yet recommend in category queries gets credited for being "known" rather than collapsing to zero. Full formula here. You can always drill into the individual metrics for a deeper breakdown.

The tool is now live at visibilitytool.ai: a simpler, faster take on the metrics described on this site. No waitlist or sign-up process needed; head over and check your brand's AI visibility directly.

No. This directory is free to browse with no account or login. To check your own brand's AI visibility, use our tool at visibilitytool.ai.

We built a three-level taxonomy: 23 top-level categories, broken into subcategories, broken into 459 specific industries. We then identified the notable brands competing in each one, just over 10,000 brands in total. Every brand was analyzed in the context of its industry, so scores always reflect how a brand performs against its actual competitive set rather than against all brands globally.

The directory is a snapshot: all scores were measured in April 2026. Language models change frequently, with new versions shipping and training data getting updated, so treat these scores as a point-in-time reference for how the brand landscape looked, not a live feed. For a current check of your own brand, use our tool at visibilitytool.ai.