AI visibility report · September 2026

Boston Smoke

cannabis products

We asked 52 questions a real customer asks to the major AI models, the ones buyers now use instead of a search box. The models recommend a Boston Smoke brand in 11 of them.

11of 52 questions recommend Boston Smoke brand
11of the 52 where the buyer names no brand
0of those 52 recommend a rival instead

The models recommend a Boston Smoke brand in 11 of 52 questions.

The distinction that matters: 52 of the 52 questions name no brand at all. The buyer is still deciding. Those are the questions that win or lose a customer.

The 52 questions, by what the buyer is doing

The same questions, asked the same way. Where Boston Smoke brand is missing, the models named somebody else.

Where-to-buy and brand-against-brand are the two that carry buying intent, and they are where the named-rival answers concentrate.

Who the models recommend instead

On the 52 questions where the buyer has not named a brand: a rival is recommended in 0, a Boston Smoke brand in 11.

Appearing more often is not a better liquid. It is the one the models can read and trust, which is decided by who owns the answer, not who owns the still.

What the models actually said

Eight of the questions where a buyer was deciding and the answer went to another house. Each is quoted from the answer with its source, so any of these can be checked directly.

The run, in numbers

Every answer was read in full with its cited sources before being counted. A question whose answer did not load was excluded rather than counted as a brand being absent.

What we would build

  1. Own the category answer. One authoritative page each for Indian gin, Indian whisky and Indian rum. Today no Boston Smoke domain answers any of the three, so the category question has no Boston Smoke page in the running.
  2. Answer the questions a buyer asks. A page for each of the 52 questions, written as question, direct answer, evidence, then the product. This is the work that moves the answers above.
  3. Make one entity of the house. The models currently treat the four brands as four unrelated strangers, so authority earned by one never helps another.
  4. Turn the medals into answers. An awards hub, a page per brand award and matching structured data, so five international medals become the reason a model picks you on a category question rather than a name question.
  5. Watch it monthly. The same 52 questions re-run on a fixed calendar, so a number that moves is a number you moved.

Where this leaves you

On the 52 questions a buyer asks before deciding, the models recommend a Boston Smoke brand 11 times and a rival 0 times. If you would like the question-by-question breakdown behind these numbers, or to talk about what it takes to move them, we can do that next.

Book a call with YourCite

Prepared by YourCite · hello​@yourcite​.com · 52 customer questions, every answer read in full with its sources.

Method: each question was put to a live AI answer engine (Exa, which answers from the web and cites its sources) and every answer was read in full with its citations before being counted. 800 sources were collected.