In 35% of answers, a business we measure is named.
The answers are about dive centres; fishing operators; DMCs; and hotels and guesthouses in Seychelles.
176 of 504 answers — the same questions put to Perplexity, ChatGPT, Claude every morning, 29 October 2026 to 12 November 2026.
- 4
- niches measured deeply enough to publish
- 25
- businesses measured in those niches
The 4 niches we measure deeply enough to publish
The floor is 5 measured businesses and 4 phrasings. A niche below it is measured and not printed.
| Niche | Measured | Named | Since |
|---|---|---|---|
| dive centres | 7 | 110 of 168 | 21 August 2026 |
| fishing operators | 7 | 0 of 84 | 28 August 2026 |
| DMCs | 6 | 49 of 126 | 21 August 2026 |
| hotels and guesthouses | 5 | 17 of 126 | 21 August 2026 |
Named: over the comparable mornings of 29 October 2026 to 12 November 2026, the days when every assistant answered.
Measured and Since: over the whole row, 21 August 2026 to 12 November 2026.
One answer counted once for each business we asked it for.
Businesses we measure, not businesses that exist — and the names of measured businesses are never published, here or anywhere.
Where the assistants get Seychelles’s names
Every page cited in answers about those niches, by share of all citations.
- tourism.seychelles.travel5.4%
- booking.com5.1%
- seyvillas.com4.8%
- padi.com4.4%
- business websites, between all of them14.7%
Tourism-board and government pages take 10.6% of every citation here. The most-cited page of all is one nobody here controls.
5 September 2026 to 12 November 2026 · 33,751 citations in 4,306 answers · the niches above
What a board sees
The sector, not a list of companies
Every business we measure in a niche, asked the same questions on the same days, reported as the state of that niche.
Where the assistants take names from
The kinds of source cited across every stored answer for a niche in a market, and the domains cited most often.
Where the board’s own site sits among those sources
Whether the official destination site is cited at all, and how often, against the catalogues and directories that are.
Share of voice beside neighbouring destinations
How often the destination is named when a traveller asks a question that names no country.
A row of days, not a snapshot
One run moves a great deal on its own. We report change only across a row of comparable days.
A second and third language as separate series
Each language is measured on its own days and never averaged with another.
A written report each quarterdone by a person
Written by a person, not generated — and marked as drafted until the board has read it.
How often any domestic business is named
When an assistant is asked where to go, a destination is either in the answer or it is not.
We measure the whole sector — every business we can find, the same questions, once a week.
And the figure that matters to a board is not one business’s but the share of answers that name anybody at home.
- 4.0%95 businesses · 14 comparable days
- 3.7%165 businesses · 14 comparable days
- 2.9%227 businesses · 14 comparable days
- 2.4%284 businesses · 14 comparable days
- 2.3%200 businesses · 14 comparable days
- 1.8%290 businesses · 14 comparable days
Dive centres, 6 markets, the same questions on the same days.
4.0% of answers name a domestic business in the strongest and 1.8% in the weakest — 2 times apart.
- 3.0%75 businesses · 14 comparable days
- 1.4%69 businesses · 14 comparable days
Fishing operators, 2 markets, the same questions on the same days.
3.0% of answers name a domestic business in the strongest and 1.4% in the weakest — 2 times apart.
The markets are not named, and neither would yours be.
11 of the 21 niche-and-market pairs we measure are not here: below five businesses or three comparable days we have nothing to compare.
A further 2 are measured in another language.
The questions there are asked as that market’s travellers ask them, and a share from a different question bank is not comparable with these.
A pilot is what fills a destination’s own row.
One number, from our own measurement
Dive operators in 6 markets, the same questions, every morning from 2026-09-01 to 2026-10-05.
10,824 readings in all — one reply counted once for each business we asked it for.
In the weakest of those markets the assistants named an operator we measure there in 0.6% of them. In the strongest, 13.9%.
Dive operators, six markets, 1 September to 5 October 2026 — a fixed reading, not a live count.
The operators we measure in that market, not every operator there.
Over the same window, some business was named in 58.7% of answers in the weakest market and 97.9% in the strongest; the ones we measure, in 2% and 84.5%.
The businesses in both are ordinary ones of the kind your members are.
hotels and chains, travel agencies, tour operators and DMCs, and the operators who sell the trips.
How a pilot works
Placed to end a month or two before your next budget is drafted.
Ninety is the length because a month can only show you the state of your sector, never what changed in it.
A change across a row of comparable days is the finding.
Where it ends is the part we fit to you.
- We write nothing in the board’s name and approach no business as though sent by it.
- Each one sees its own measurement and decides for itself.Nothing about an individual business goes to the board.
- The niches, the sources, the share of voice, and the row of days behind each figure.
After it, and what cannot be caught up
Buying does not start the measurement and declining does not stop it.
The same questions go to the assistants once a week either way; what ends with the pilot is your access to the row.
The part nobody can undo later is the reading you did not take.
A gap is never filled in afterwards, so months spent without access stay missing from your own record permanently.
Talk to us
What you send reaches team@almanet.ai and is not used for anything else. How we handle it.