Where is it worth being for AI to find you? Slovak directories and forums vs global platforms
Of the 79,571 links AI models used while answering questions from the Slovak market, 5,500 led to a page a company can actually do something about — either sign up to it, or watch what other people are writing there. Whether those forums and business directories are Slovak is decided mostly by sector: in IT a Slovak domain shows up in only 7 % of such answers, in every other sector in 60.6 %.
What we measured
The sample is 8,331 answers from ChatGPT, Perplexity and Gemini to questions from the Slovak market (in Slovak and for a Slovak location) between July 1 and August 31, 2026. 6,800 of them cited at least one link — which, at 79,571 citations, means an average of 11.7 links per answer.
A citation is one URL in one answer. What a citation means matters too: it says the model searched the web while answering and leaned on that page. It says nothing about what it learned during training — a missing domain is therefore no proof that it played no part in the answer. We wrote about how those cited pages hold up technically in a separate analysis.
The tracked brands are not evenly represented in the sample. Wherever we give a percentage, it is a share within that sector, category or model — never a comparison of market sizes.
Every cited URL is sorted in citeme[eu] into one of seven categories (competitor, directory, community, reference, marketplace, other, or your own site). This analysis is about two of them: the business directory and the community page, i.e. the forum.
Directories vs. forums: where to register and what to watch
A directory is a page where a company creates its profile, or where the profile is created by the provider. That covers g2.com, clutch.co and trustpilot.com, but equally the catalog on azet.sk or zoznam.sk and a company profile on profesia.sk. What they have in common is that you can sign up: fill in the form, add the description, ask for reviews if you like. The profile is created by the company, not user.
A forum page is one somebody else wrote about the company: a Reddit thread, a question on porada.sk, a discussion on modrykonik.sk, a user post on LinkedIn. What you can do there is follow the posts and join the discussion. The six pairs below show what that looks like in the actual data:
A company can sign up to it, fill it in and keep it current. The model then reads it and uses it as a description of the business.
„kde kúpiť starožitný porcelán online“
https://www.zoznam.sk/katalog/Domacnost-zahrada-kancelaria/Starozitnosti„kde riešiť zmluvy a firemnú agendu v Žiline“
https://www.azet.sk/katalog/notari/zilina„firmy na vývoj podnikového softvéru na mieru“
https://clutch.co/sk/developersThere is nothing to sign up to. All you can do is know it exists, read it and join the conversation.
„hľadám starožitné obrazy“
https://www.modrykonik.sk/market/starozitnosti-obrazy„POS terminál na festivaly a trhy“
https://www.modrykonik.sk/forum/nakupne-poradenstvo/ktoreho-dodavatela-terminalu-na-prevadzku-odporucate/„najlepšie chaty pre skupinový pobyt pri lyžiarskom svahu“
https://www.reddit.com/r/Slovakia/comments/18kgalhIn our sample directories account for 3,928 citations — 4.9 % of all 79,571 — appearing in 22.7 % of the answers that leaned on web sources, which is 18.6 % of all 8,331 answers. Forums account for 1,572 citations, only 2 % of all citations, but appear in 14.6 % of web-grounded answers — 12 % of all of them. Between them, at least one of the two shows up in 32.9 % of the answers that leaned on web sources, or 26.8 % of every answer.
Some platforms belong to both. We counted LinkedIn 523 times as a directory and 151 times as a forum — the same domain, with nothing but the URL path to separate them. The same goes for Facebook: the business page is a directory, the group is a forum.
| Domain | Directory | Forum | Total |
|---|---|---|---|
| linkedin.com/company and /in vs /posts and /pulse | 52378 % | 15122 % | 674 |
| facebook.combusiness pages vs /groups | 9985 % | 1815 % | 117 |
The Slovak-to-foreign split depends on the sector
The chart below focuses on the answers from AI chatbots that cited a directory or a forum. Within those answers we then measured the share held by Slovak platforms. How far Slovak directories and forums make it into an answer depends on the sector. The reason is commercial, not technical. A question like "recommend a project management tool" is a software buying question, and the model breaks it into searches that land on G2, Clutch or Product Hunt. A Slovak equivalent of those platforms barely exists. A law firm or an e-shop, by contrast, gets looked up in a catalog that does exist in Slovak — and the models do use it. Accommodation sits somewhere in the middle, since alongside platforms like TripAdvisor there are Slovak alternatives such as kamnavylet.sk.
How often a sector's answers reach a Slovak (.sk) domain — separately for directories and forums
How to read this: of the answers to e-shop questions where the model cited at least one directory, 88.3 % contained at least one Slovak directory. In IT it is 7.2 %. This is not a share of links but a share of answers.
The results depend not only on the sector but also on which AI model or chatbot is used. In IT all three models are effectively level. That is no coincidence: where no Slovak platform exists, there is nothing for a model to prefer. Where a Slovak alternative does exist, the models diverge sharply. It is not the case, though, that one model is simply "more Slovak" than the others: in fintech it is ChatGPT that cites Slovak sources most often, and in accommodation it is Perplexity.
The main thing behind those differences is the language the model searches in. Answers for which the AI model broke the user's question into English searches reached a Slovak domain in only 2.7 % of cases; those that searched in Slovak, in 69.3 %. English searching happens mostly on IT questions — we took apart how that fan-out language is chosen and what it means for Slovak companies in the previous article. The table below shows the share in each sector separately for ChatGPT, Gemini and Perplexity — though for simplicity directories and forums are counted together.
Slovak share of directories and forums by sector and model
| Sector | ChatGPT | Gemini | Perplexity |
|---|---|---|---|
| Law | 73.8 % | 100 % | 93 % |
| E-shop | 43.2 % * | 82.4 % | 92.1 % |
| FinTech | 66.1 % | 3.3 % * | 50.4 % |
| Accommodation | 1.4 % | 6.1 % * | 61.7 % |
| IT & technology | 9.2 % | 3 % | 5.5 % |
How to read this: when Perplexity answered an e-shop question and cited a directory or a forum, a Slovak domain was among them in 92.1 % of cases. For ChatGPT it was 43.2 %. We always compare within a single sector — otherwise the numbers would compare questions rather than models.
* Fewer than 50 answers in the cell — a single answer moves the percentage by several points, so read it as a direction, not a number.
So which business directories matter most — where should you spend time and attention?
Answering that across the board is hard, because the results depend on the sector as well as on the model. For this analysis we ordered them by the citeme[eu] score, a number derived from the percentage of answers a domain appears in. The score is a weighted average across 5 separate sectors. This ordering serves only as a recommendation of which sites to look at first.
Directories and review platforms by citeme.eu score.
634 citations · 373 answers
593 citations · 372 answers
99 citations · 83 answers
523 citations · 289 answers
90 citations · 82 answers
236 citations · 126 answers
669 citations · 310 answers
107 citations · 46 answers
275 citations · 150 answers
256 citations · 150 answers
How to read this: azet.sk has a score of 23.6 — meaning it appears in that percentage of the answers where the model cited any directory at all, once the sectors are averaged with equal weight. It is not a share of links: when a model lists eleven subpages of the same site in one answer, that counts as one answer. Scores therefore do not add up to 100 %.
† A domain whose pages fall into different categories depending on the URL path. The row is labelled with the root domain even though a specific section was cited — blog.sme.sk and diskusia.sme.sk, for instance, not the newsroom on sme.sk.
Which forums and communities should you watch?
As with the business directories, we used the citeme[eu] score here too. Without converting to a score, platforms such as GitHub would sit higher, since their importance to the IT sector is considerable.
Community sites by citeme.eu score.
696 citations · 472 answers
141 citations · 114 answers
282 citations · 205 answers
151 citations · 121 answers
71 citations · 63 answers
20 citations · 20 answers
18 citations · 16 answers
19 citations · 18 answers
89 citations · 67 answers
44 citations · 42 answers
How to read this: reddit.com has a score of 39.7 — meaning it appears in that percentage of the answers where the model cited any forum at all. As with directories this is a share of answers, not of links, and scores do not add up to 100 %.
† A domain whose pages fall into different categories depending on the URL path. The row is labelled with the root domain even though a specific section was cited — blog.sme.sk and diskusia.sme.sk, for instance, not the newsroom on sme.sk.
What to do about it
In the order we would tackle it:
- Register in the Slovak business directories the models draw on: the Azet catalog and the Zoznam catalog, the two main platforms. Add zlatestranky.sk and profesia.sk if you are planning to hire.
- Fill in your LinkedIn company page and your Google profile (Google Business Profile). Both behave as directories in the data, both show up in every sector, and both are entirely under your control.
- If you sell software or development services, G2 and Clutch matter a great deal — in IT they are the most-cited directories, and there is no Slovak substitute for them.
- If you run accommodation or a restaurant, deal with TripAdvisor and your Facebook business page — the two most-cited platforms in that sector. This article does not cover platforms like booking.com or megaubytovanie.sk — but for accommodation in Slovakia they matter a great deal.
- Watch reddit, modrykonik.sk, porada.sk and the like, depending on your sector. citeme[eu] helps here: it records when AI models cite your company or products negatively on the basis of posts from those platforms.
The advantage is that none of this is expensive or complicated. Most of it is one afternoon's work, and in the data it shows up as the difference between the model finding and mentioning you in an answer, or not. From the entry plan up, citeme[eu] shows which of these pages the models cite for your brand specifically, on your questions.