AI models name Singapore Airlines most often, and four audited carriers never appear in an answer

When a buyer asks an AI model which airline to fly, the answers go to a small group of carriers. Singapore Airlines is named in 35 of 60 recommendation moments. Delta Air Lines is named in 30. Ethiopian Airlines, Iberia, SAS and Swiss are named in none.

Scores run from 86 to 54, with a median of 67. Singapore Airlines is the only carrier in the Awesome tier. Swiss, SAS, Iberia, Lufthansa and Korean Air sit in the Weak tier.

In this category, brand strength and AI recommendation often disagree. Turkish Airlines equals Singapore Airlines on Sentiment, but the models name it in only 4 of 60 moments. Emirates scores 10 on Findability, yet the models name it in 26 of 60 moments.

That gap is where the opportunity is. Many carriers have real equity that the models do not yet reward. A smaller group gets recommended above its equity, and that position may be harder to keep.

ABOUT THIS READING

ARA scores each brand on five attributes, each out of 20: Findability, Clarity, Sentiment, Identity Fidelity and AI Recommendation. The total puts each brand in a tier. AWESOME is 83 to 100. STRONG is 70 to 82. AVERAGE is 56 to 69. WEAK is 40 to 55. INVISIBLE is 0 to 39.

ARA asked 4 AI models 15 buyer questions about 30 airlines. That gives 60 recommendation moments for each brand. This is the first reading of the category, so there is no prior quarter to compare against. Each brand also gets a position label. SWELL means real equity that the machines under-reward. EBB means the machines recommend the brand above its equity. ALIGNED means the two agree.

The questions state no market, so each model chose its own frame. GPT answered 7 of 15 questions with US airlines or US data. Claude did this on 4 questions, Perplexity on 4 and Gemini on 3. The other answers ranked airlines worldwide. No model assumed a market other than the US. Claude named no airline on 1 question, the question about trust. GPT, Perplexity and Gemini named airlines on all 15 questions. ARA counts a decline as a finding, not as a missing answer.

Winners and losers

Singapore Airlines ranks first at 86, the only Awesome score. All 4 models name it. It appears in 35 of 60 moments and is named first in 7. It wins the general airline question, long-haul comfort, safest airline, best service and best business class. It is absent from cheapest fare and best loyalty programme. Its lowest attribute is Findability, at 13.

Delta Air Lines scores the maximum on AI Recommendation, against a cohort average of 8.8. The maximum score does not make Delta the most-named carrier. The models name Delta first in 7 of 60 moments, the same as Singapore Airlines, and name it in 30. Delta wins most trusted, best value, reliability and recovery, business traveller preference and the general airline question. It is absent only from cheapest fare.

Swiss ranks last at 54. No model names Swiss in any of the 60 moments. That includes the general airline question, most trusted, safest airline and best service. Swiss scores 16 on Clarity and 14 on Sentiment. Those fundamentals do not reach the answers.

Ethiopian Airlines scores 0 on AI Recommendation, against a cohort average of 8.8. The models never name it. Its Clarity is 18, the same as Singapore Airlines and Delta Air Lines. Its Sentiment is 16. That makes Ethiopian Airlines one of the clearest Swell cases on the board.

The Leaderboard.

The audited carriers hold 75.7% of the airline shelf

Across all answers, the models made 334 choices from 67 brands. The 30 audited carriers hold 75.7% of those choices. The largest single holding is 33 choices, or 9.9%. No one airline owns the shelf.

The biggest names outside the audited set are EVA Air with 7 choices, Frontier Airlines and Wizz Air with 5 each, and AirAsia, Allegiant Air, Malaysia Airlines and STARLUX Airlines with 4 each. Several of these are low-cost carriers. The cheapest fare question is the one that Singapore Airlines and Delta Air Lines do not win.

The tail is long. The models name 18 brands exactly once, for 5% of all choices. They name 13 brands exactly twice, for 8%. Together those 31 brands hold 13% of the shelf. Switch the view to Full Shelf to see them beside the audited set.

15 buyer questions x 4 models / audited brands in colour, every other brand in a stepped grey. Toggle between Cohort Only and Full Shelf.

KEY FINDING 01

Turkish Airlines equals Singapore Airlines on Sentiment and appears in 4 of 60 answers

Turkish Airlines scores 18 on Sentiment and 18 on Clarity. Singapore Airlines has the same two scores. The models name Singapore Airlines in 35 moments and Turkish Airlines in 4. Turkish Airlines is never named first.

The same pattern repeats across the Swell group. Air France scores 16 on Sentiment and is named in 4 moments. Virgin Atlantic has the highest Identity Fidelity on the board, at 18, and is named in 3. Air New Zealand totals 71 and is named in 7.

For these carriers, the equity exists. The answers do not reflect it yet. The gap between the two is the work.

KEY FINDING 02

Qatar Airways, Emirates and Etihad all get recommended above their equity

All three Gulf carriers hold the Ebb position. Emirates scores 10 on Findability, among the lowest on the board, but the models name it in 26 of 60 moments. Qatar Airways scores 11 on Findability and is named in 25.

Three large US carriers share the position. United Airlines has the lowest Identity Fidelity on the board, at 11, and still scores 17 on AI Recommendation. American Airlines has the lowest Sentiment, at 12, and scores 18 on AI Recommendation. Alaska Airlines is also in the Ebb position.

An Ebb position is a lead the fundamentals do not fully support. If the models start to weight equity more closely, these carriers may have the most to lose. The dot-grid below plots each carrier's fundamentals against its recommendation.

Brand Fundamentals vs Recommendation.

KEY FINDING 03

Etihad is named first in 5 of its 10 appearances, and Emirates in 1 of 26

Being named and being named first are different results. Emirates appears in 26 moments but leads only 1. Etihad appears in 10 moments and leads 5.

Southwest Airlines leads 4 of its 23 appearances. Cathay Pacific leads 4 of 21. Qatar Airways leads 4 of 25. A carrier that appears often but rarely leads is on the list without being the answer.

KEY FINDING 04

The models understand most airlines, and fewer airlines read as distinct

Clarity scores are high across the board. Singapore Airlines, Delta Air Lines, United Airlines, Qatar Airways, LATAM and Ethiopian Airlines all score 18. Lufthansa is the low point, at 13.

Identity Fidelity varies more. United Airlines scores 18 on Clarity and 11 on Identity Fidelity. The models know what United is, but they reproduce less of what makes it United. Virgin Atlantic shows the reverse case. It scores 18 on Identity Fidelity and 16 on Clarity.

Singapore Airlines scores high on both, with 18 on Clarity and 17 on Identity Fidelity. It is also the most-named carrier. The chart below places each carrier on the two axes.

Understood vs Distinct.

KEY FINDING 05

No airline scores above 15 on Findability

JetBlue has the top Findability score, at 15. easyJet, Lufthansa and SAS score 9. Even Singapore Airlines scores only 13.

Findability is the ceiling for the whole category. Carriers in the Swell group need the models to find their equity before the models can recommend it. A carrier that raises Findability first may gain the most ground per point of effort.

WHAT THIS MEANS

Singapore Airlines and Delta Air Lines hold the top of the answers. Each one still has open questions. Neither wins cheapest fare. Singapore Airlines is also absent from best loyalty programme. Those are the lanes where the leaders can still gain ground.

For the Swell carriers, the fundamentals already exist. Turkish Airlines, Ethiopian Airlines, Air France and Virgin Atlantic have strong Clarity, Sentiment or Identity Fidelity scores, but the models rarely name them. This is the first reading of the category. The next quarter will show which carriers close the gap between equity and recommendation.