In This Article
Five Danish government pages across four domains share a skeleton and a blind spot. AI can read what each page calls itself. The audited pages give it almost no structured evidence about who is in charge, which authority is responsible, or when anything was last updated. That gap matters more than usual in 2026. The Mette Frederiksen III government, which calls itself the four-leaf-clover government, has governed Denmark since 3 June 2026. Digitalisation was folded into a combined research, education and digitalisation ministry the same day. The institutions changed. The machine-readable record of them did not.
In This Article
- National average: 47.2, Grade D. Denmark ranks fifth of the six national profiles so far, ahead of Croatia (40.4). The range is 41 to 54. No page reaches Grade C.
- regeringen.dk leads at 54, one point short of Grade C. The digital minister’s page sits at 41. The gap is not content. It is dates, entities and schema.
- Still zero JSON-LD on all five. Entity connectivity, knowledge-graph anchoring and E-E-A-T density are 0/100 on every page.
- Four of five have a broken H1 structure. em.dk carries 14 H1 tags, ism.dk and regeringen.dk carry 2 each, and digst.dk has none. Only the minister’s page raised no H1 alert.
- Freshness is thin. Only regeringen.dk (61) and digst.dk (57) have a detectable date, and neither is in JSON-LD. The other three score 8, 4 and 0.
- The two pages that should name the people in power name nobody. The Inspector detects no primary entity on regeringen.dk or on the digital minister’s page. Entity graph stability is 15/100 on both.
- Denmark has a clear AI Act authority, and none of it is marked up. Digitaliseringsstyrelsen was designated Denmark’s national supervisory authority under the EU AI regulation, with a coordinating role across Danish authorities and towards the EU AI Office. Its homepage has no H1 and no schema.
The profile in one view
Scores are the Inspector’s dimension scores (0 to 100) for each page, ordered by overall index.
| Area | regeringen.dk | digst.dk | ism.dk | em.dk | regeringen.dk minister page |
|---|---|---|---|---|---|
| Content architecture | 70 | 60 | 70 | 65 | 75 |
| Data extractability | 67 | 80 | 59 | 65 | 62 |
| Entity clarity | 45 | 45 | 45 | 45 | 45 |
| Schema and metadata | 25 | 25 | 35 | 20 | 10 |
| Freshness signals | 61 | 57 | 4 | 8 | 0 |
| Trust and authorship | 0 | 0 | 0 | 0 | 0 |
| Overall index | 54 (D) | 53 (D) | 45 (D) | 43 (D) | 41 (D) |
Shading: 75 and above strong, 50 to 74 moderate, 20 to 49 weak, below 20 absent. The pages are regeringen.dk, digst.dk, ism.dk, em.dk and the regeringen.dk profile page for Christina Egelund, the minister who holds the digital portfolio.
Denmark is the reverse of Czechia. Czechia has strong architecture (91.0) and weak metadata and freshness (18.0 and 10.6). Denmark has the second-best schema (23.0) and freshness (26.0) scores after Cyprus, but the weakest architecture of the five non-Cypriot countries (68.0). The Danish headline gaps are not the tags themselves. They are discipline: one H1 per page, one date per document, one entity per institution.
The shared absence is the same as everywhere else in the series: zero JSON-LD, zero trust signals, and no machine-readable identity for the institutions that run national AI policy.
How the five engines read the Danish pages
| Engine | regeringen.dk | digst.dk | ism.dk | em.dk | Minister page | Average |
|---|---|---|---|---|---|---|
| Claude | 62 | 61 | 60 | 59 | 63 | 61.0 |
| ChatGPT | 61 | 64 | 58 | 59 | 61 | 60.6 |
| Copilot | 43 | 44 | 48 | 50 | 46 | 46.2 |
| Perplexity | 52 | 53 | 41 | 38 | 34 | 43.6 |
| Gemini | 44 | 42 | 37 | 31 | 27 | 36.2 |
Engines that reward readable prose and hierarchy (Claude, ChatGPT) are tolerable on Danish pages. Engines that lean on structured data, freshness and authorship (Gemini, Perplexity) are weak. Gemini averages 36.2 and falls to 27 on the minister’s page. Heading disorder costs less than the reports’ own warnings suggest. em.dk, with 14 H1 tags, scores 59 on Claude. The minister’s page, with no H1 alert, scores 63.
Where Denmark needs to improve
Ranked by how many pages are affected and how much one fix would unlock.
1. Structured identity (all five)
No Organization, Person or Article schema exists on any page. The Danish AI Act framework is clear on paper. The supplementary AI Act law designates Digitaliseringsstyrelsen, Datatilsynet and Domstolsstyrelsen as national competent authorities. Digitaliseringsstyrelsen is the notifying authority, the central contact point and a market surveillance authority for prohibited AI practices. In October 2025 it published a package of six guidance documents on the AI practices the regulation prohibits. None of this relationship, between the agency, its two co-authorities and its guidance, exists as a connected entity on digst.dk. A machine sees an organization name and nothing else.
2. The people at the top are invisible (regeringen.dk and the minister’s page)
The Inspector detects no primary entity on either page. Entity graph stability is 15/100, labelled unstable. The Prime Minister’s Office site does not register the Prime Minister. The page for the minister who holds the digital portfolio does not register the minister. Christina Egelund was appointed on 3 June 2026 to a ministry that merged the former education and research ministry with the digitalisation ministry. Its underlying bodies include Digitaliseringsstyrelsen, Danmarks Statistik and CPR. That reporting line is exactly the kind of relationship Person and Organization markup exists to express.
3. Dates that exist for humans and not for machines (all five)
The government’s political platform page carries a plain-text last-updated date of 16 June 2026. That is a date a person can read. The Inspector credits regeringen.dk and digst.dk with date detection through text or meta tags, and both keep their freshness score below 62 because nothing sits in JSON-LD. The other three have no machine-readable date at all. ism.dk scores 4, em.dk 8 and the minister’s page 0. These are pages announcing a new government, new ministers and new policy, and engines that weigh recency cannot tell they are recent.
4. Broken H1 structures (four of five)
digst.dk has no H1, which the Inspector flags as a missing document anchor. em.dk carries 14 H1 tags. ism.dk and regeringen.dk carry two each. The minister’s page raised no H1 alert and posts the best structure score of the five (75). The Danish fix is mechanical. One H1 naming the institution, with everything else demoted to H2 or H3, would move four pages at once.
5. A government in transition is the cheapest moment to fix this
The new ministry structure means Danish government sites are being rebuilt or renamed. ism.dk says so itself: it announces that a new joint website for the health and church ministry is being prepared. Identity markup is cheap to add to a build in progress and expensive to retrofit. Ministry names, ministers, start dates and reporting lines are all fresh facts this year. They are also the ones engines will be asked about.
Institution profiles
regeringen.dk – Grade D – 54/100
The Prime Minister’s Office site, one point from Grade C, with no entity AI can name
Strengths: The highest overall score in the set and the best freshness score (61). Dates are detected through text or meta tags. It publishes the political platform of the four-party government. Its best-scoring candidate query, at 55% eligibility, is the one about that platform.
What AI sees: “Regeringen” as the page topic, two H1 tags and no primary entities. Entity graph stability is 15/100. Four images lack alt text (56% coverage) and no canonical URL is credited. Of its 15 candidate queries, several are interface text: “Gå til indhold” (skip to content), “Skift cookie indstillinger” (change cookie settings) and “Læs mere her” (read more here). The Inspector also classified a headline about a military donation package for Ukraine as an AI query, which tells you how little topical signal the page gives it.
Missed opportunity: A GovernmentOrganization record for the Prime Minister’s Office, Person markup for the Prime Minister and the ministers the site lists, and Article markup with datePublished and dateModified for the platform page. That page already states its own update date in plain text. The markup would turn it into a signal.
digst.dk – Grade D – 53/100
The AI Act authority’s homepage has no H1
Strengths: The best data extractability in the set (80) and the strongest ChatGPT reading (64). It is the only page where the Inspector detects “Artificial Intelligence” as a primary concept, at 83% clarity. Its headlines include a notice that the first supervision of prohibited AI practices has concluded.
What AI sees: “Digitaliseringsstyrelsen” as the page topic, but with no H1 to anchor it. The only entities are “Artificial Intelligence,” a generic “Product / Service” and a generic “Organization / Brand,” with no schema. Its 23 candidate queries average 40.9% eligibility, the lowest in the set. Several are navigation text: “Gå til indhold,” “Skift cookie indstillinger,” “Genveje” (shortcuts) and “Ofte besøgte sider” (frequently visited pages). Three images lack alt text and no canonical is credited.
Missed opportunity: This is the page that should be the machine-readable root of Danish AI Act supervision. A single H1, an Organization record with parentOrganization pointing to the combined ministry and sameAs links to Datatilsynet and Domstolsstyrelsen, and Article markup for the prohibited-practices guidance with publication dates would let engines cite digst.dk as the authority it is by statute. The role is already in law. The markup is not.
ism.dk – Grade D – 45/100
A ministry in transition, with the best schema score and almost no dates
Strengths: The highest query eligibility in the set (51.1% on average, up to 54%) and the best schema and metadata score (35). Entity graph stability is 48, the highest of the five. It is the only page where a named person registers: the Inspector detects the new minister, Ida Auken, as an emerging concept.
What AI sees: “Sundheds” as the page topic, a truncated first word of the ministry’s name. It has two H1 tags, 475 words and no machine-readable date (freshness 4). The Inspector also tags a fragment, “Most Favoured,” as an emerging concept, which looks like a tokenization artefact. The page announces the new government, the minister’s appointment and a joint website under construction.
Missed opportunity: The site is being rebuilt, so the identity layer can go in at build time. That means Organization markup for the health and church ministry, Person markup for the minister with role and start date, and NewsArticle markup with dates for the headlines already on the page: the AMR high-level representative, the fertility inspection supplement, the acute plan for child and adolescent psychiatry, and the cancer after-effects guidance. The content is dated by the world. It is undated by the page.
em.dk – Grade D – 43/100
Fourteen H1 tags and a page AI calls “Front page”
Strengths: The strongest Copilot reading in the set (50) and some of the most specific policy content: green investment support, rules for substantiating green claims, a removal of 1.3 million illegal products from an online platform, and the Danish role in EU burden-reduction work. All of it is written for people.
What AI sees: “Forside” (front page) as the page topic and 14 H1 tags. There is no meta description and 13 images lack alt text, which leaves coverage at 13%, the worst in the set. The only detected entity is a generic “Product / Service.” Freshness is 8. Gemini scores it 31, the lowest of any Danish homepage.
Missed opportunity: The 14 H1s are most likely each news teaser marked up as a top-level heading, which should be confirmed in the page source. One H1 naming the ministry, with the teasers demoted, plus a meta description and NewsArticle markup with dates, would address the three biggest costs at once. The page also lists the agencies and state-owned companies under the ministry. That is already an organizational hierarchy waiting for markup.
regeringen.dk minister page (Christina Egelund) – Grade D – 41/100
The digital minister’s page: no entity, no date, no schema
Strengths: The best structure score of the five (75), the highest ChatGPT citation estimate in the set (about 47%) and the highest-scoring query anywhere in the sample: 59% eligibility for the minister’s title and name. She was appointed on 3 June 2026 to the merged research, education and digitalisation ministry. The page has a CV and personal financial interests section, which is exactly the kind of authorship and accountability content engines look for.
What AI sees: “Forsknings” as the page topic, the first word of a long ministerial title. It detects no primary entities and no dates (freshness 0). Schema is 10, there is no meta description, no canonical and three images lack alt text (40% coverage). Gemini scores it 27 and Perplexity 34, the lowest in the set. At least five of its ten candidate queries are navigation text, including “Gå til indhold,” “Om regeringen.dk,” “AKTUELT” and “MINISTRENE.”
Missed opportunity: This page should be a Person record: name, jobTitle, worksFor (the combined ministry), startDate of 3 June 2026, sameAs links to the official ministry and parliament profiles, and a dateModified. It would also connect the minister to the agency that supervises the AI Act. Right now the page is the weakest link in a chain of responsibility Denmark has otherwise defined clearly.
Scope and method
Each page was audited once on 6 October 2026 with the AI Visibility Inspector (v1.9.3) and the Ivica Srncevic Framework. This is a snapshot of five pages across four domains, not a full-site audit. Four are homepages. The fifth is the regeringen.dk profile page for the minister with the digital portfolio. If it were excluded, the four-homepage average would be 48.8. All pages were audited in Danish. The identical entity clarity score of 45 on all five is a feature of the scoring model rather than of the sites, since the Inspector reports no JSON-LD anywhere. This profile treats schema as absent and reads that credit as coming from page-level metadata and entity tokenization. Query categories and engine citation percentages are the tool’s estimates and are heuristic. The “content gaps” list is identical on all five and was left out. Facts about the Danish government and AI Act framework come from public sources checked on 6 October 2026. The Inspector measures what a page exposes, not what an institution does.
The national average of 47.2 is the arithmetic mean of the five overall index scores: 54, 53, 45, 43 and 41.
This research series also includes National AI profiles of Austria, Belgium, Croatia, Cyprus, Czechia and Denmark.
Read also Denmark Flagship Companies.
Calculate your potential exposure below
Inputs
Enter your own data where possible. Hover over labels or open “Methodology” for details on how each field is used.
From GA4, GSC, or your SEO platform. Use a 30–90 day average.
Session → lead / signup / purchase. Use decimal (e.g. 1.2 for 1.2%).
For B2B, use average deal value or LTV. Currency is symbolic here.
Used to show profit‑based cost of inaction. Leave at 100 to ignore.
Share of organic traffic from queries/pages likely to be influenced by AI answers and retrieval systems.
Based on observed traffic loss patterns in AI search and structural decay cases.
Expected organic growth if you do nothing special about AI visibility.
Plausible uplift over the chosen horizon based on case studies.
How these numbers are calculated (methodology)
Let:
- O₀ = monthly organic sessions
- CVRₒ = organic conversion rate (as a decimal)
- ARPC = average revenue per conversion
- GM = gross margin (as a decimal)
- E = AI exposure share (as a decimal)
- D = displacement rate (as a decimal)
- T = time horizon in months
Baseline monthly revenue:
- Revenue: \(R_0 = O_0 imes CVR_o imes ARPC\)
- Profit: \(R_{0,profit} = R_0 imes GM\)
Monthly revenue at risk (for a given D):
- \(R_{risk} = R_0 imes E imes D\)
- Profit version: \(R_{risk,profit} = R_{risk} imes GM\)
The tool computes R_risk for three scenarios:
- Conservative: D = 0.15
- Base: D = 0.30 (or your selected scenario)
- Aggressive: D = 0.50
Cost of Inaction over T months is approximated as:
- \(CoI(T) = R_{risk} imes T\)
- Shown as a range from conservative to aggressive.
The growth inputs (g₀, g₁) are currently used for narrative context and can be incorporated into a more advanced version that models opportunity cost explicitly.
