In This Article
Estonia is advancing one of Europe’s most coherent AI policy frameworks – but the five official pages that should make that framework visible to AI systems are almost invisible to retrieval engines.
An AI Visibility Inspector audit of five Estonian government homepages across three domains finds a national average of 48.6/100, Grade D. All five pages cluster between 48 and 49, with perfect heading structure and near-zero structured identity. The institutions responsible for Estonia’s third national AI strategy (2024–2026), its Data and AI White Paper (2024–2030), the KrattAI vision for intelligent public services and the new Eesti.ai productivity programme all score in the same narrow band of weakness.
In This Article
- National average: 48.6, Grade D. Estonia ranks in the middle of the National AI Profiles series so far, with all five audited pages between 48 and 49. No page reaches Grade C.
- Two institutions, three domains. The audits cover the Government of Estonia (valitsus.ee/en), the Ministry of Economic Affairs and Communications (mkm.ee/en) and the Information System Authority (ria.ee/en). Valitsus.ee and mkm.ee each appear twice in the five-row set due to re-audits.
- Structure is strong everywhere. Metadata is absent everywhere. All five pages score 100 on structural integrity, but schema and metadata scores are 10/100 and freshness signals are 0/100 on every page.
- Still zero JSON-LD on all five. Entity connectivity, knowledge-graph anchoring and E-E-A-T density are at or near 0/100 on every page.
- The pages that should name Estonia’s AI direction name almost nothing. The Inspector detects weak or unstable entity graphs on all five homepages, despite their role in publishing national AI policy.
The profile in one view
The table below shows the five audited institutions (rows) and the five dimension scores that determine the overall index. Scores are from the AI Visibility Inspector (v1.9.3), audited on 6 October 2026.
| Institution (URL) | Structural integrity | Data extractability | Entity clarity | Schema and metadata | Freshness signals | Overall index |
|---|---|---|---|---|---|---|
| valitsus.ee/en (audit 1) | 100 | 75 | 40 | 10 | 0 | 49 (D) |
| mkm.ee/en (audit 1) | 100 | 70 | 40 | 10 | 0 | 48 (D) |
| ria.ee/en | 100 | 75 | 40 | 10 | 0 | 49 (D) |
| valitsus.ee/en (audit 2) | 100 | 75 | 40 | 10 | 0 | 49 (D) |
| mkm.ee/en (audit 2) | 100 | 70 | 40 | 10 | 0 | 48 (D) |
Shading convention (for your CMS): 75 and above strong, 50 to 74 moderate, 20 to 49 weak, below 20 absent.
Estonia’s 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, almost no trust signals, and no machine-readable identity for the institutions that run national AI policy.
How the five engines read the Estonian pages
The Inspector reports compatibility scores for five engines: Perplexity, OpenAI/ChatGPT, Claude, Google Gemini and Microsoft Copilot. The table below shows the five institutions (rows) and the five engine scores (columns).
| Institution (URL) | Perplexity | ChatGPT | Claude | Gemini | Copilot |
|---|---|---|---|---|---|
| valitsus.ee/en (audit 1) | 27 | 72 | 76 | 19 | 58 |
| mkm.ee/en (audit 1) | 27 | 70 | 75 | 14 | 58 |
| ria.ee/en | 27 | 72 | 76 | 19 | 58 |
| valitsus.ee/en (audit 2) | 27 | 72 | 76 | 19 | 58 |
| mkm.ee/en (audit 2) | 27 | 70 | 75 | 14 | 58 |
Engines that reward readable prose and hierarchy (Claude, ChatGPT) are tolerable on Estonian pages. Engines that lean on structured data, freshness and authorship (Gemini, Perplexity) are weak. Gemini averages 17.0 across the five rows and falls to 14 on mkm.ee audit. Perplexity is flat at 27 on every page. The structural strength of the pages does not compensate for missing schema, dates and entity signals.
Where Estonia 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 of the five audited pages. The AI Visibility Inspector reports entity connectivity and knowledge-graph anchoring at 0/100 on all valitsus.ee and mkm.ee audits, and near-zero on ria.ee.
Estonia’s AI policy landscape is clear on paper. The Action Plan for Artificial Intelligence 2024–2026 was created by the Ministry of Economic Affairs and Communications, the Ministry of Justice and the Ministry of Education and Research. The Data and Artificial Intelligence White Paper 2024–2030 sets the longer-term direction. KrattAI defines the vision for intelligent public services, and Eesti.ai frames AI as a productivity and economic-growth instrument.
None of this relationship, between the ministries, agencies and programmes, exists as a connected entity on the audited homepages. A machine sees an organization name and almost nothing else.
2. The people and institutions at the top are invisible (all five)
The Inspector detects no primary entities on valitsus.ee/en and mkm.ee/en, and only weak entity graphs on ria.ee/en. Entity graph stability scores are 15/100 (unstable) on valitsus.ee, 41/100 (fragmented) on mkm.ee and 53/100 (fragmented) on ria.ee.
The Government Office homepage does not register the Prime Minister or the government as a named entity. The ministry homepage does not register the minister responsible for digital affairs or AI policy. The Information System Authority page detects “Artificial Intelligence”, “Technology” and “Author / Expert” as concepts, but without Person or Organization schema these remain unanchored tokens.
These are exactly the relationships that Person and Organization markup exists to express: which ministry leads AI policy, which agency implements it, which ministers are responsible and when their mandates began.
3. Dates that exist for humans and not for machines (all five)
All five pages score 0/100 on freshness signals. The Inspector reports no JSON-LD dateModified, no machine-readable publication or modification date, and no content age under 180 days on any page.
These are pages that publish national strategies, white papers and AI programmes with specific time horizons (2024–2026, 2024–2030). Engines that weigh recency cannot tell they are recent. The audits explicitly flag “No date signals found – content age is unverifiable. Add dateModified JSON-LD.” on all five pages.
A plain-text “last updated” line is not enough. Perplexity, Gemini and other retrieval-driven systems look for datePublished and dateModified in JSON-LD Article or similar schema.
4. Missing Article, FAQ and author schema (all five)
All five pages lack Article or FAQPage JSON-LD, and none has Person schema for authors or responsible officials. The audits label this a high-priority issue on every page.
The consequences are direct:
- No Article schema means AI systems cannot anchor citations to a clear content type.
- No Person schema means E-E-A-T is “entirely unverifiable” for engines like Gemini.
- No FAQ schema means question-shaped content (for example, “What is KrattAI?” or “What is Eesti.ai?”) does not register as FAQ content.
The audits also note missing or too-short meta descriptions on all five pages, which reduces the quality of LLM-generated summaries used in citations.
5. A government in active AI expansion is the cheapest moment to fix this
Estonia is in a phase of active AI policy development and public-sector implementation. The Government Office launched the Eesti.ai initiative in early 2026, with a stated goal of doubling the value of work in Estonia by 2035 and growing GDP by 25% within five years. In 2026, €10.98 million was directed to Eesti.ai to increase AI skills, public-sector efficiency, healthcare services and education.
This is the optimal moment to add identity markup. Ministry names, ministers, strategy dates and reporting lines are all fresh facts this year. They are also the ones engines will be asked about when users query “Estonia national AI strategy”, “KrattAI” or “Eesti.ai”. Identity markup is cheap to add to a build in progress and expensive to retrofit.
Institution profiles
valitsus.ee/en (audit 1) – Grade D – 49/100
The Government of Estonia homepage, with strong structure and no machine-readable identity
Strengths: Perfect structural integrity (100) and solid data extractability (75). The page publishes information about the government, its ministers and its development strategy, including “Estonia 2035”. Its best-scoring candidate queries include “Prime Minister, Ministers News and contacts ‘Estonia 2035’ development strategy” at 49% eligibility.
What AI sees: “Avaleht” as the page topic, a single H1 and logical heading nesting, but no primary entities. Entity graph stability is 15/100, labelled unstable. One image lacks alt text (partial coverage), and no JSON-LD schema is detected. Of its 19 candidate queries, several are generic or interface-related: “What is avaleht?”, “What is accessibility?”, “What is contrast?” and cookie/accessibility notices.
Missed opportunity: A GovernmentOrganization record for the Government of Estonia, Person markup for the Prime Minister and relevant ministers, and Article markup with datePublished and dateModified for key policy pages. The page already references the “Estonia 2035” strategy and government news. The markup would turn those references into citable entities.
mkm.ee/en (audit 1) – Grade D – 48/100
The AI policy ministry’s homepage, with the weakest engine scores in the set
Strengths: Perfect structural integrity (100) and adequate data extractability (70). The page publishes news about entrepreneurship, innovation, work and equal opportunities, and cooperation agreements relevant to economic and digital policy. Its strongest candidate queries reach 54% eligibility.
What AI sees: “Home page” as the page topic, a single H1 and logical nesting, but only generic “Organization / Brand” and “Product / Service” entities, with no schema. Entity graph stability is 41/100, labelled fragmented. Five images lack alt text (72% coverage), and no JSON-LD schema is detected. Gemini scores it 14, the lowest of any Estonian homepage in this profile.
Missed opportunity: This is the page that should be the machine-readable root of Estonia’s AI and data policy. An Organization record for the Ministry of Economic Affairs and Communications, with parentOrganization pointing to the Government of Estonia and sameAs links to relevant EU and national registers, plus Article or NewsArticle markup with dates for strategy and news content, would let engines cite mkm.ee as the authority it is by statute. The role is already in law and policy documents. The markup is not.
ria.ee/en – Grade D – 49/100
The Information System Authority homepage, with the strongest entity signals and no schema
Strengths: Perfect structural integrity (100) and the best entity clarity in the set, with “Artificial Intelligence”, “Technology”, “Author / Expert” and “Product / Service” detected as entities. Entity graph stability is 53/100, the highest of the five audits. The page includes content on cyber security, AI and state information systems, and its candidate queries average higher eligibility than the other two pages.
What AI sees: “Home” as the page topic, a single H1 and logical nesting, but still no JSON-LD schema. Schema and metadata score is 10/100 and freshness signals are 0/100. The Inspector flags high-priority issues: no Article or FAQ schema, no Person schema, and no machine-readable dates.
Missed opportunity: This is the agency that operates Bürokratt and leads the Aruait innovation project for secure public-sector AI. An Organization record for RIA, with clear relationships to the Ministry of Justice and Digital Affairs and to KrattAI/Eesti.ai, plus Article markup for AI and cyber security content with publication and modification dates, would make ria.ee the natural citation source for questions about Estonia’s public-sector AI and cyber security.
valitsus.ee/en (audit 2) – Grade D – 49/100
The Government of Estonia homepage, re-audited, with identical weaknesses
Strengths: Identical to audit 1: structural integrity 100, data extractability 75, entity clarity 40, schema and metadata 10, freshness 0. The page continues to publish government news, ministerial information and strategy content.
What AI sees: Identical to audit 1: “Avaleht” as topic, no primary entities, entity graph stability 15/100 (unstable), no JSON-LD schema, and a similar set of generic and policy-related candidate queries.
Missed opportunity: Identical to audit 1. The re-audit confirms that the structural strengths and metadata weaknesses are stable over time, not a one-off configuration issue.
mkm.ee/en (audit 2) – Grade D – 48/100
The AI policy ministry’s homepage, re-audited, with identical weaknesses
Strengths: Identical to audit 1: structural integrity 100, data extractability 70, entity clarity 40, schema and metadata 10, freshness 0. The page continues to publish news on entrepreneurship, innovation, work and equal opportunities.
What AI sees: Identical to audit 1: “Home page” as topic, generic organization and product entities without schema, entity graph stability 41/100 (fragmented), five images missing alt text, and no JSON-LD schema. Gemini again scores 14.
Missed opportunity: Identical to audit 1. The re-audit confirms that the ministry’s role as the statutory lead for AI policy is not reflected in machine-readable identity.
Scope and method
Each page was audited once on 6 October 2026 with the AI Visibility Inspector (v1.9.3) and the Srna SEO Framework. This is a snapshot of five page audits across three domains, not a full-site audit. All five are English-language homepages for institutions directly involved in Estonia’s AI policy and implementation.
If it were excluded, the four-audit average for the remaining pages would be 48.75. All pages were audited in English. The identical entity clarity score of 40 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. Facts about the Estonian government and AI strategy 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 48.6 is the arithmetic mean of the five overall index scores: 49, 48, 49, 49 and 48.
This research series also includes National AI profiles of Austria, Belgium, Croatia, Cyprus, Czechia and Denmark.
Read also Estonia 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.
