You invest in SEO, you rank well on Google — but when someone asks ChatGPT for the best solution in your category, your name is not there. How do you know how invisible you are? And how do you measure whether that changes?
With SEO you measure rankings — how high you appear in Google for a keyword. That is a relatively stable measurement: pages rank based on links, authority and relevance.
AI visibility measures something different: how often and how positively your brand is recommended when someone asks a question to an AI model. There is no Google Search Console for ChatGPT. No dashboard that tells you how often Gemini has recommended you.
Without systematic measurement you work blind — you do not know what you are missing and cannot track whether your approach is working.
AI visibility vs SEO
Build a list of fifteen to twenty discovery queries and category comparison queries in both Dutch and English. Focus on questions potential customers ask before they know your brand.
Open ChatGPT, Gemini, Perplexity and Claude separately. Enter each query in a new conversation and note per model and language: is your brand mentioned, at what position, with what sentiment?
Interpret results across five dimensions: visibility score, discovery score, niche authority score, competitor intercept score and sentiment score. Each metric points to a specific action.
Dutch-language and English-language queries produce structurally different results — even for the same brand in the same category. Bilingual tracking shows where your biggest gap is.
Manual measurement works for a baseline. For structural optimisation it is not sustainable. Use an automated platform to run the same queries weekly and compare scores with the previous period.
How often does your brand appear in answers, expressed as a percentage across all queries, models and languages. The starting point for all other metrics.
Are you found on exploratory questions, or only when someone already knows you exist? Low = invisible to new customers.
How strongly do AI models associate you with your specific category? High = seen as the logical player in your market.
How often are you recommended as an alternative when someone asks about a competitor? One of the most valuable positions to occupy.
When you are mentioned, in what tone? Recommended, neutral, or with reservations? Negative sentiment is rare but damaging.
This is where most brands go wrong: they measure only in English, or only in Dutch. The results are structurally different.
A brand prominently recommended in English-language ChatGPT answers can be completely absent from Dutch-language answers about the same category — and vice versa. Training data differs per language, and models pick up different sources depending on the query language.
For Dutch brands specifically: Dutch-language training data is scarce. With relatively little Dutch content you can already build a strong position — but only if you know where you currently stand. Bilingual measurement shows where your biggest gap is and in which language you can gain the most ground fastest.
Algora measures
Weekly is the minimum if you are actively optimising. Monthly is acceptable for maintenance. One measurement per quarter gives no usable trend data — you can see where you stand, but not what caused any changes.
LLM answers vary per session, but patterns are definitely measurable. By running each query multiple times and averaging the results, you get a reliable picture. Automation with multiple runs per query gives a more stable signal than one-off manual checks.
Results are input for your content strategy. A low discovery score means too little structured content that answers exploratory questions. A low competitor intercept score means missing comparison pages or mentions on authority platforms.
Ideally yes. ChatGPT has the largest user share. Perplexity searches live and cites sources — visibility there is more directly influenced. Gemini is relevant for users in the Google ecosystem. Claude is relevant in professional contexts. Measuring all four gives a complete picture.
Then the content may not yet be indexed, or may not be the type that AI models cite. HowTo pages and FAQs with schema markup are picked up faster than generic blog posts. If scores do not move after six to eight weeks, the strategy needs reviewing.
AI models are already answering questions from your potential customers. The only question is whether your brand is part of the answer.
Start your free AI visibility audit →~90 seconds · no credit card required