Rankfor.AI, the leader in AI visibility, presents the first peer-reviewed AI reputation ranking spanning the Nordic, Baltic, and Central European markets. 11 countries, 12 languages, 3 leading AI models.

The Rankfor® Index 2026 · Nordic-Baltic Edition

How AI sees 66 European brands. In 12 languages. At the same time.

35,640 AI responses. 66 brands across 11 markets. 12 languages. 3 AI models. One composite score per brand, per language. Open data, peer-reviewed methodology, free for the industry.

Get the full report →Browse all 66 brand profiles

brands

languages

markets

35,640

AI responses

The Rankfor® Index 2026 Leaderboard.

60+ Leaders50-60 Middle<50 Development

Rank	Brand	Tier	Sector	Citations	Blind spots	Rankfor® Index score
🥇 1	Spotify	Global	Tech	725	0	66.5
🥈 2	Evolution	Global	Tech	464	0	66.2
🥉 3	Kone	Global	Industrial	501	0	65.9
4	Wise	Global	Fintech	940	0	65.7
5	Printful	Global	Tech	1178	0	65.5
6	Vestas	Global	Energy	803	0	65.4
7	Maersk	Global	Logistics	702	0	65.4
8	Lidl Slovakia	Global	Retail	529	0	64.5
9	Zalando	Pan-European	Retail	485	0	64.5
10	Pipedrive	Global	Tech	882	0	63.9
11	Klarna	Global	Fintech	768	0	63.6
12	Wolt	Pan-European	Tech	741	0	63.3
13	IKEA	Global	Retail	699	0	62.3
14	Bolt	Global	Tech	570	0	62.1
15	H&M	Global	Retail	665	0	62.0
16	Slovak Telekom	National	Telecom	566	1	61.8
17	Vinted	Global	Tech	556	0	61.3
18	Nokia	Global	Tech	543	0	60.5
19	Carlsberg	Global	Consumer goods	628	1	58.3
20	Volkswagen Slovakia	Global	Automotive	641	1	58.0
21	Volkswagen	Global	Automotive	510	1	57.7
22	SAP	Global	Tech	582	0	57.3
23	Volvo Cars	Global	Automotive	793	1	56.9
24	Lego	Global	Consumer goods	644	1	55.4
25	PKN Orlen	Pan-European	Energy	587	4	54.3
26	Swedbank Estonia	Nordic-Baltic	Banking	875	5	52.6
27	Allianz	Global	Insurance	553	4	52.2
28	Tele2 Lithuania	Nordic-Baltic	Telecom	814	5	52.1
29	Swedbank Latvia	Nordic-Baltic	Banking	921	5	52.1
30	Deutsche Telekom	Global	Telecom	790	7	51.8
31	Swedbank	Nordic-Baltic	Banking	586	5	51.8
32	Kahoot!	Global	Tech	581	2	51.7
33	Vipps	Nordic-Baltic	Fintech	590	4	51.7
34	ESET	Global	Tech	883	4	51.2
35	Telenor	Pan-European	Telecom	593	5	51.2
36	Supercell	Global	Tech	778	4	51.2
37	Novo Nordisk	Global	Pharma	555	5	51.0
38	Equinor	Global	Energy	654	4	50.9
39	Allegro	Pan-European	Retail	462	6	50.4
40	REWE Group	Pan-European	Retail	851	4	50.4
41	Skoda Auto	Global	Automotive	695	4	50.2
42	Kesko	Nordic-Baltic	Retail	695	4	49.6
43	Reserved	Pan-European	Retail	548	6	49.5
44	Elisa Estonia	Nordic-Baltic	Telecom	689	5	49.5
45	airBaltic	Nordic-Baltic	Aviation	761	6	48.9
46	Statkraft	Pan-European	Energy	810	5	48.8
47	Elisa	Nordic-Baltic	Telecom	435	5	48.8
48	InPost	Pan-European	Logistics	489	6	48.5
49	Danske Bank	Nordic-Baltic	Banking	734	8	48.4
50	Pilsner Urquell	Global	Consumer goods	621	6	48.0
51	Rimi Baltic	Nordic-Baltic	Retail	713	7	47.9
52	CD Projekt	Global	Tech	746	7	47.1
53	DNB	Nordic-Baltic	Banking	500	5	47.1
54	Avast	Global	Tech	640	5	47.0
55	Lekaren Dr. Max	National	Retail	472	8	47.0
56	Alza.cz	National	Retail	456	7	46.9
57	Maxima	Nordic-Baltic	Retail	359	7	46.9
58	Kiwi.com	Pan-European	Tech	652	6	46.8
59	Luminor Bank	Nordic-Baltic	Banking	872	9	46.7
60	mBank	National	Banking	527	7	45.6
61	Ignitis Group	Nordic-Baltic	Energy	732	9	45.6
62	CSOB	National	Banking	616	9	45.5
63	Lietuvos paštas	National	Logistics	731	9	44.5
64	Selver	National	Retail	365	9	43.3
65	Latvijas Gāze	National	Energy	1010	10	43.1
66	Tatra Banka	National	Banking	1006	10	42.9

Source: Rankfor.AI Rankfor® Index 2026, Nordic-Baltic Edition. Peer-reviewed dataset, 35,640 AI responses, May 2026.

See all 66 brand profiles →

Cross-country vertical winners.

Which brand wins each industry across all 11 Nordic-Baltic markets. Rankings combine sentiment, source quality, cross-language consistency, and recommendation share.

#	Brand	Market	Tier	Rec. share	Mean ARI
1	Volkswagen Slovakia	Slovakia	global	65%	58.0
2	Volkswagen	Germany	global	64%	57.7
3	Volvo Cars	Sweden	global	57%	56.9
4	Skoda Auto	Czech Republic	global	33%	50.2

Country-by-country top brands.

Top brands per home market, scored using the brand's home language plus English. Tells you who AI defaults to when the buyer is in-market.

#	Brand	Industry	Rec. share	Mean ARI
1	Spotify	Tech	100%	66.4
2	H&M	Retail	100%	66.2
3	Volvo Cars	Automotive	92%	65.7
4	Klarna	Fintech	100%	65.5
5	IKEA	Retail	92%	64.1
6	Swedbank	Banking	50%	53.2

The Bilingual Penalty.

How much does AI shift its view of a brand between its home language and English? The biggest finding: global multinationals get penalized at home, local champions get boosted at home.

Penalized at home

AI is more positive in English than in the brand's home language.

Skoda Auto

Czech 47.0 · English 52.7

-5.7

DNB

Norwegian 44.1 · English 49.1

-5.0

Volvo Cars

Swedish 47.3 · English 51.5

-4.2

Volkswagen Slovakia

German 46.9 · English 51.1

-4.2

Statkraft

Norwegian 47.0 · English 50.5

-3.5

Swedbank

Swedish 43.8 · English 47.3

-3.5

Lidl Slovakia

German 46.0 · English 49.5

-3.5

Deutsche Telekom

German 46.3 · English 49.4

-3.1

Swedbank Estonia

Swedish 44.7 · English 47.7

-3.0

Maxima

Lithuanian 46.3 · English 48.9

-2.6

Boosted at home

AI is more positive in the brand's home language than in English.

Vinted

Lithuanian 49.9 · English 40.6

+9.3

CD Projekt

Polish 48.9 · English 40.1

+8.8

ESET

Slovak 48.1 · English 40.4

+7.7

Avast

Czech 46.7 · English 39.8

+6.9

Bolt

Estonian 50.2 · English 44.5

+5.7

Kahoot!

Norwegian 49.9 · English 44.4

+5.5

Novo Nordisk

Danish 49.8 · English 44.6

+5.2

InPost

Polish 47.7 · English 42.8

+4.9

Kiwi.com

Czech 46.9 · English 42.0

+4.9

Pilsner Urquell

Czech 46.8 · English 42.0

+4.8

Nordic-Baltic + CEE 2026

The AI link economy.

131,510 citations across 21,059 unique domains. Global HHI 0.0035 — atomistic, no domain dominates. The top-1 domain (wikipedia.org) captures only 4.4% of all citations.

The 80/20 rule barely holds: 80% of all citations come from 3,878 domains — that's 18.4% of the 21,077 unique domains AI grounds in. The remaining 20% of citations require 17,000+ more domains to cover. AI's evidence base is genuinely long-tailed.

Cumulative distribution of cited domains across the merged Nordic-Baltic + CEE 2026 AI Reputation Index. Top-1 domain captures 4.41% of citations; top-10 captures 11.2%; top-50 captures 24.5%; top-1000 captures 62.3%. 80% of citations come from 3,878 domains (18.4% of the 21,077 unique domains). — Cumulative distribution of cited domains. The first 50 domains capture 24.6% of citations; reaching 62.3% requires 1,000 domains. AI grounds in a long tail.

Total citations

131,510

Unique domains

21,059

Global HHI

0.0035

Top-50 share

24.6%

LanguageThe language of the AI promptEach row aggregates every citation AI returned when asked about any of the 66 brands in this language.Lithuanian = the row covers all citations made in Lithuanian, across all brands and all models.	CitationsHow many times AI cited any sourceTotal number of citation links AI returned in this slice. One AI response can cite many sources.If GPT cites BBC, Reuters, and Wikipedia in one answer, that's 3 citations.	Unique domainsHow many different websites AI usedDistinct domains that appeared at least once. The breadth of AI's evidence base.17,891 unique domains overall — essentially every credible source on the web.	HHIConcentration index (0–1)Herfindahl-Hirschman index. Sum of squared citation shares per domain. Lower = AI grounds in many balanced sources. Higher = a few domains dominate.HHI 0.005 = atomistic (no domain controls the narrative). HHI 0.05 = noticeably concentrated. HHI 0.15+ = a single source dominates.	Top-1 domainThe most-cited single source in this sliceThe domain that AI cited most often. Often Wikipedia for general slices; can shift to local press in narrow ones.In Lithuanian, the top-1 is vz.lt (local business press), not Wikipedia. Localization matters.	Top-1 %Share of all citations from the top-1 domainWhat percentage of every citation in this slice comes from a single most-cited source.5% means even the most popular domain accounts for 1 in 20 citations. 25%+ means one source dominates the slice.
English	11,628	3,460	0.0047	wikipedia.org	4.8%
Polish	11,437	3,232	0.0036	wikipedia.org	3.7%
Slovak	11,310	3,210	0.0034	wikipedia.org	3.8%
Czech	11,307	3,005	0.0043	wikipedia.org	4.5%
German	11,239	3,220	0.0045	wikipedia.org	4.4%
Swedish	11,046	3,174	0.0041	wikipedia.org	4.4%
Latvian	10,736	2,684	0.0055	wikipedia.org	4.6%
Estonian	10,608	2,726	0.0065	wikipedia.org	4.4%
Finnish	10,580	2,802	0.0065	wikipedia.org	5.7%
Lithuanian	10,576	2,735	0.0062	vz.lt	4.4%
Norwegian	10,523	3,051	0.0051	wikipedia.org	4.9%
Danish	10,520	3,147	0.0039	wikipedia.org	4.0%

HHI = Herfindahl-Hirschman index of citation share by domain. Lower = more fragmented (good for diverse PR coverage). Higher = a few sources dominate (controllable through targeted partnerships). HHI > 0.05 indicates meaningful concentration; > 0.10 indicates a single-source bias risk.

Five things the data says.

66.5

Spotify leads the Nordic-Baltic AI cohort.

Spotify wins with mean ARI 66.5 across 12 languages, followed by Evolution (66.2), Kone (65.9), Wise (65.7), Printful (65.6), Vestas (65.4), and Maersk (65.4). 18 brands score above 60. The new top-10 is dominated by AI-monopoly tech and industrial brands rather than legacy automotive leaders.

+23.1

Local champions get a massive home-language boost.

Reserved gains 23.1 ARI points when AI is asked in Polish vs English. airBaltic gains 22.9 in Latvian. Tele2 Lithuania, Swedbank Latvia, Lietuvos paštas, and Elisa each gain 22+ points at home. AI defaults to mentioning these brands locally, but barely recognizes them when speaking English.

−7.6

Deutsche Telekom is most penalized at home.

Deutsche Telekom loses 7.6 ARI points in German vs English. Pipedrive (−4.0 in Estonian), VW Slovakia (−3.8 in German), Lidl Slovakia (−3.4), Kahoot! (−2.9 in Norwegian), and Allianz (−2.8) follow. Domestic German press is consistently the most critical of established multinationals.

97%

Wise, Kone, and Vestas own their AI categories.

Wise (Fintech, 97% rec share), Kone (Industrial, 97%), Vestas (Energy, 96%), and Maersk (Logistics, 94%) are AI-locks: when AI is asked about the category in Nordic-Baltic, these brands appear in nearly every response. Spotify also hits 100% rec share for Tech-Sweden. Banking, by contrast, has no AI default: the leader (Swedbank Estonia) commands only 46%.

21,077

No domain controls the AI link economy.

Citations span 21,077 unique domains across the merged Nordic + Baltic + CEE 2026 dataset. Wikipedia is the largest single source at 4.41% of all citations. Long-tail dominates: top 50 domains capture only 24.5%; reaching 80% requires several thousand domains. Owned-media still outperforms paid for individual brands.

Or in plain English.

Imagine your buyer walks into three different libraries (ChatGPT, Gemini, Perplexity) and asks each librarian the same question about your brand. Each library has books in twelve languages. Four things happen, and we score all four.

Hover any brand in the leaderboard above to see this brand's actual numbers.

25% of the score

Does the librarian say nice things?

When the librarian describes your brand, are the words warm or cold? "Trusted, reliable, popular" pushes the score up. "Controversial, declining, criticized" pushes it down. We measure this with a multilingual sentiment model.

Plain example: A librarian saying "Lego makes the most loved toys in the world" scores higher than "Lego is one of many toy brands."

25% of the score

Does the librarian recommend you when asked?

If a buyer asks the librarian "Who are the top fintech companies in Europe?", does your brand make the list? The more often AI mentions your brand by name when buyers ask the category question, the higher the score. This is the AI default we all want to be.

Plain example: Wise is mentioned in 97% of "top fintech" responses → big points. A bank that AI never mentions when buyers ask about banks → almost nothing on this dimension.

20% of the score

Where did the librarian get the story?

Is the librarian quoting the New York Times, or a random blog? Tier-1 news, academic papers, and government sources count for a lot. Wikipedia counts a little. "I just kind of know" counts the least.

Plain example: If your AI story is sourced from BBC, Reuters, FT, your score goes up. If AI says nice things but only quotes random forums, your score goes down.

20% of the score

Is the story the same in every language?

If you ask in Lithuanian, Polish, German and English, do you get the same brand story? Or four different stories? AI that tells one consistent narrative across borders earns a higher score than AI that tells contradictory stories.

Plain example: "Volvo is famous for safety" in 12 languages = high score. "Volvo is a Swedish car company" in 6 languages but "Volvo struggled in 2024" in the other 6 = low score.

10% of the score

Do the three librarians agree?

ChatGPT, Gemini, and Perplexity all describe your brand. If they tell roughly the same story, you have a stable AI reputation. If they wildly disagree, your buyer's experience depends on which AI they happen to use.

Plain example: ChatGPT says "Bolt is a fast-growing ride-hailing leader." Gemini says "Bolt is a regional player." Perplexity says "Bolt is a controversial gig company." Your buyer gets three different stories. Score drops.

The math, in one line

Take 25 cents of niceness, 25 cents of recommendation share, 20 cents of source quality, 20 cents of consistency, and 10 cents of stability. Add them up. That's your dollar of AI reputation, on a scale of 0 to 100.

For the technical reader.
The full methodology.

The Rankfor® Index is a composite of four components, each independently validated. Every weight is published. Every formula is reproducible. The underlying dataset has been submitted to Springer Discover Artificial Intelligence for peer review.

25%

Sentiment

Multilingual sentiment analysis (XLM-RoBERTa) of every AI response, normalized to 0-100. Captures how positively AI describes the brand.

20%

Source Quality

Weighted average of citation source types. Tier-1 news and academic rank highest, implicit knowledge lowest. Measures the quality of evidence AI grounds in.

20%

Consistency

Cross-language semantic alignment (BGE-M3 embeddings). How similar the brand narrative is across 12 languages.

25%

Recommendation Share

Percentage of buyer-intent prompts (F1 + E1) where AI mentions the brand by name when asked the category question. Measures whether AI defaults to recommending the brand.

10%

Stability

Inter-model agreement across GPT, Gemini, and Perplexity. Brands with high stability tell the same AI story regardless of which model is asked.

Rankfor® Index v3 = 0.25 × Sentiment + 0.20 × Source Quality + 0.20 × Consistency + 0.25 × Recommendation Share + 0.10 × Stability

Get the report

Get the full report.

29 slides + recommendation audit. In your inbox in minutes. Free.

For press and analysts.
No form required.

Journalists, researchers, and analysts can access the methodology and press release directly. No email, no form. Industry research is meant to move.

PDF · Whitepaper

Methodology whitepaper

Full formula, source type weights, statistical tests, reproducibility notes, and legal framework.

PDF · Press release

Press release

One-page media release with headline findings, top 10 table, quotes, and contact details.

Direct request

Press kit + charts

High-resolution charts, headshots, and interview availability. Email Dmitrij directly.

Right of reply

Your brand is in the index?
You have the right of reply.

Every brand in the Rankfor® Index 2026 is offered the opportunity to submit a response. Responses are published alongside the brand's profile. If you represent a brand in the ranking, email Dmitrij with subject line "Right of reply" followed by your brand name.

Submit a response →

About the Rankfor® Index. The Rankfor® Index 2026 measures cross-language AI reputation at a single point in time (April 2026). It is a composite diagnostic score, not a brand quality ranking. Low scores indicate opportunity for AI reputation work, not failure of the underlying business. Every brand in the index has a distinctive strength, highlighted in the full report.

Methodology. Full methodology, weights, source type classifications, and reproducibility notes are published in the methodology whitepaper. The underlying dataset has been submitted to Springer Discover Artificial Intelligence for peer review.

Citation. Żatuchin, D. (2026). The Rankfor® Index 2026: Nordic-Baltic Edition. Rankfor.AI Research Report. Retrieved from open.rankfor.ai/index-2026

How AI sees 66 European brands. In 12 languages. At the same time.

The Rankfor® Index 2026 Leaderboard.

Cross-country vertical winners.

Country-by-country top brands.

The Bilingual Penalty.

Penalized at home

Boosted at home

The AI link economy.

Five things the data says.

Or in plain English.

Does the librarian say nice things?

Does the librarian recommend you when asked?

Where did the librarian get the story?

Is the story the same in every language?

Do the three librarians agree?

For the technical reader.The full methodology.

Get the full report.

For press and analysts.No form required.

Your brand is in the index?You have the right of reply.

We use cookies

For the technical reader.
The full methodology.

For press and analysts.
No form required.

Your brand is in the index?
You have the right of reply.