Seedlip takes the most first places in non-alc spirits, while AI names Lyre's most often

Seedlip leads the Non-Alc Spirits board at 79. AI names it first in 6 of 60 recommendation moments, more than any other brand. Lyre's sits one point behind at 78. AI names Lyre's in 23 of 60 moments, the highest count in the cohort.

The top of the board is crowded. Seedlip, Lyre's, Ritual Zero Proof, Ghia, Pathfinder and Monday all score strong. The median is 60. Scores run from 79 down to 49.

The larger opportunity sits lower on the board. Six brands carry more equity than AI rewards. Three Spirit and Kin Euphorics post the highest Identity Fidelity in the cohort, 17 each. AI names Three Spirit in 6 of 60 moments and Kin Euphorics in 4.

AI also gives part of the category to brands outside the cohort. The 15 brands hold 47.7% of the shelf. Non-alcoholic beers such as Athletic Brewing and Guinness 0.0 appear in the answers.

ABOUT THIS READING

ARA asked four AI models 15 buyer questions, for 60 recommendation moments. ARA scored 15 brands on Findability, Clarity, Sentiment, Identity Fidelity and AI Recommendation. Each attribute is out of 20, and the total is out of 100. The tiers are AWESOME 83 to 100, STRONG 70 to 82, AVERAGE 56 to 69, WEAK 40 to 55 and INVISIBLE 0 to 39. A SWELL brand has real equity that AI under-rewards. An EBB brand gets recommended above its equity.

This is a repeat reading, and the instrument changed between quarters. Last quarter read Claude Sonnet 4.6, GPT-4o, Perplexity sonar-pro and Gemini 2.5 Flash. This quarter reads Claude Sonnet 5, GPT-5 search, Perplexity sonar-pro and Gemini 3.5 Flash. Part of any movement on the eight returning brands comes from that model change, not from the brands. Last quarter asked each question twice per model and averaged the two answers. This quarter asks each question once, so part of any change in AI Recommendation may be sampling.

The sampling change does not affect the per-brand answers. ARA scored Identity Fidelity on a revised rubric this quarter, with Brand Character removed and Brand Messaging added. Identity Fidelity movement is therefore not like-for-like, and a reader must not treat it as brand movement on its own. ARA measured movement on the same 15 questions as last quarter. The cohort grew from 8 to 15 brands. Six of the eight returning brands have a comparable reading, and the seven new brands have a first reading only.

Pathfinder and Lapo's report no movement. Last quarter's prompts asked about the bare names. Many answers described another entity, such as the Nissan Pathfinder or Lapo Elkann, or came from a model with no web access. Pathfinder scored 7 and Lapo's 17 then, against 75 and 49 now. Those differences correct the earlier reading. They are not growth, and the cohort median leaves both brands out.

Lapo's name still sends answers elsewhere. In 28 of 64 answers this quarter, a model described another entity, such as Lapo Elkann or a restaurant, or asked which Lapo's the buyer meant. That confusion is a finding about the brand, not a defect in the reading.

Winners and losers

Seedlip holds the top of the board at 79, down 1 point since last quarter. Findability fell 2 points and Identity Fidelity rose 2. ARA scored Identity Fidelity on a revised rubric, so that rise is not like-for-like. Part of the movement comes from the model change. AI chooses Seedlip 18 times in 60 moments, across all four AI models. Seedlip is absent from the questions on wine replacement, dinner parties, winding down and a drink that makes you feel something.

Ghia posts the largest rise, up 15 points to 77. Clarity rose from 10 to 16, and Identity Fidelity rose from 12 to 16. The revised rubric means the Identity Fidelity rise is not like-for-like. Part of the movement comes from the model change. AI chooses Ghia 20 times in 60 moments, across all four AI models.

Last quarter ARA set Ghia three bets. The Disambiguation Bet targeted Findability: lead copy and schema with the apéritif category and separate the brand from the car. ARA found no evidence of that work, and the grade is No action found. Findability rose one point, because the car and Ford collision still caps it. The Open-the-Door Bet targeted AI Recommendation through a serve-and-swap content layer. AI Recommendation rose four points, but ARA found no content layer, so the grade is Moved, unattributed.

The One-Voice Bet targeted the brand's voice, which Identity Fidelity measures. Voice rose four points, and models now quote the brand's own lines. The brand's reputation did not move, because complaints about bitterness and price persist. The grade is Partly landed, and the revised rubric limits how much of that rise counts. The tie stops at Clarity. No bet targeted Clarity, yet Clarity gave the largest part of the rise, from 10 to 16.

Lapo's sits at the bottom at 49. Last quarter's score of 17 described mostly other entities, so ARA re-baselined the brand and gives it no movement figure. AI chooses Lapo's once in 60 moments, from one of four AI models. It wins no question. In 28 of 64 answers, a model described another entity, such as Lapo Elkann or a restaurant, or asked which Lapo's the buyer meant.

Ritual Zero Proof takes the biggest-fall slot, but its score rose. It gained 3 points to 77, the smallest gain among the returning brands. Part of that gain comes from the model change. AI chooses Ritual 20 times in 60 moments, across all four AI models. Ritual is absent from the questions on wine replacement, dinner parties, New York bartenders and winding down.

ARA set Ritual three bets last quarter. The Filing Bet targeted Findability: fix ownership and category copy, bridge the tequila synonyms and set "Ritual Zero Proof" as the one canonical name. The site still splits Tequila and Agave Spirit naming, and Findability did not move. The grade is Not acted on. The RTD Attribution Bet asked Ritual to put "Zero Proof" on every can mention. The brand issued no can news, and the Margarita can page now reads "Coming Back Soon", which no model knows.

The RTD Attribution grade is Gap widened. The Coverage Bet targeted AI Recommendation with a corrected range map and cans aimed at occasion questions. ARA found no range map, and one model still invents two SKUs. AI Recommendation rose by only 1 point, and the grade is Not acted on. The tie stops there. No bet landed, so the bets do not explain the 3-point gain.

The Leaderboard.

The cohort holds less than half of what AI puts on the shelf

AI made 235 choices across 95 brands. The 15 cohort brands hold 47.7% of those choices. The largest single holding is 22 choices, or 9.4% of the shelf. No single brand holds a tenth of the shelf.

Some of the biggest names outside the cohort are beers. Athletic Brewing holds 6 choices, and Guinness 0.0 and Heineken 0.0 hold 5 each. St. Agrestis and Töst also hold 5 each. All The Bitter holds 4, and Almave Blanco and Angostura hold 3 each. AI answers some of these buyer questions with a non-alcoholic beer.

The long tail is wide. 65 brands are named exactly once, for 28% of all choices. 12 brands are named exactly twice, for 10%. Together, 77 of the 95 brands hold 38% of the shelf. AI never names Curious Elixirs or Little Saints, at any position.

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

Lyre's is the one brand AI recommends above its equity

AI names Lyre's in 23 of 60 moments, the most of any brand. Lyre's scores 19 on AI Recommendation, level with Seedlip. Its Findability is 12 and its Identity Fidelity is 14. ARA reads Lyre's as EBB: AI recommends it above the equity those scores show.

A high AI Recommendation score measures presence. First place is a separate count. AI names Seedlip first in 6 moments and Lyre's first in 5. Seedlip leads on primacy, and Lyre's leads on reach.

An EBB position can reverse when AI weighs equity more heavily. Lyre's has the reach. The brand's identity scores do not yet support it.

KEY FINDING 02

Three Spirit and Kin Euphorics carry the cohort's strongest identity and almost no recommendation

Three Spirit and Kin Euphorics each score 17 on Identity Fidelity, the highest in the cohort. Three Spirit scores 5 on AI Recommendation and Kin Euphorics scores 3. AI names Three Spirit in 6 moments and Kin Euphorics in 4. AI names neither brand first.

ARA reads six brands as SWELL: Three Spirit, Kin Euphorics, De Soi, Curious Elixirs, Little Saints and Everleaf. Each one has fundamentals that AI does not convert into recommendation. De Soi scores 15 on Identity Fidelity and AI names it once. Curious Elixirs scores 18 on Clarity and AI never names it.

ARA scored all 15 brands on the same revised Identity Fidelity rubric this quarter. The comparison holds inside the quarter. It does not hold against last quarter.

Brand Fundamentals vs Recommendation.

KEY FINDING 03

AI understands almost every brand in the cohort and tells few of them apart

Clarity runs from 14 for Lapo's to 19 for Pathfinder. Identity Fidelity runs lower, from 12 for Little Saints to 17 for Three Spirit and Kin Euphorics. AI knows what most of these brands sell. It reads less of what makes each one different.

Ritual Zero Proof scores 18 on Clarity and 13 on Identity Fidelity. Curious Elixirs scores 18 and 13. Little Saints scores 18 and 12. AI understands all three brands, and it never names Curious Elixirs or Little Saints.

Understanding alone does not earn a recommendation. A brand that AI understands but cannot tell apart gives AI no reason to choose it over the next name.

Understood vs Distinct.

KEY FINDING 04

None of the four leaderboard brands appears in answers about wine replacement or winding down

Seedlip, Ghia, Ritual Zero Proof and Lapo's are all absent from the best wine replacement question. All four are also absent from the question on a drink that helps you wind down. Seedlip, Ritual and Lapo's are absent from the dinner party question too.

The leaders win the substitution questions. Seedlip and Ritual win on sober curiosity, cutting back and going to a bar without feeling left out. Ghia wins different ground: a drink that makes you feel something, and a mom who wants to feel alive again.

These answers are not empty. Other brands fill them, and some of those brands may sit outside the cohort. A cohort brand that wants these occasions must displace whoever AI names there now.

KEY FINDING 05

Lapo's shares its name with Lapo Elkann, and AI often answers about him

In 28 of 64 answers this quarter, a model described another entity, such as Lapo Elkann or a restaurant, or asked which Lapo's the buyer meant. Lapo's scores 10 on Findability and 1 on AI Recommendation. The name confusion is a finding about the brand, not a defect in the reading.

Pathfinder shows what a clean reading can reveal. Last quarter many answers described the Nissan Pathfinder, and the brand scored 7. This quarter Pathfinder scores 75, with 19 on Clarity, the highest in the cohort. That difference corrects the earlier reading and is not growth.

Pathfinder still trails on recommendation. It scores 12 on AI Recommendation, and AI names it in 11 moments, first in one. For Pathfinder and Lapo's, the name sets the first limit on what AI can find.

WHAT THIS MEANS

The top of the board is close. Seedlip, Lyre's, Ritual Zero Proof and Ghia score between 77 and 79. At that range, first-place counts separate the leaders more clearly than totals do. Seedlip takes 6 first places, Lyre's 5, and Ritual and Ghia 2 each.

The SWELL brands hold the clearest opportunity. Three Spirit, Kin Euphorics, De Soi, Curious Elixirs, Little Saints and Everleaf carry equity that AI does not reward. AI never names Curious Elixirs or Little Saints at all. The questions on wine replacement and winding down give these brands a place to compete, because the four leaderboard brands are absent there.