The direct answer
A consultant for how hotels are found and described by AI assistants
The guest experience now starts before a property's website is opened: a traveler asks an AI assistant where to stay, and a handful of names come back. SignalTide measures whether an independent hotel or boutique spa in Miami is one of them, fixes what machines read, and reports the movement every month.
Shortlist Management is $3,000 a month on a six-month minimum for one property. Michael Brady runs a fixed set of thirty-six prompts across ChatGPT, Gemini, Perplexity, Claude and Google AI Overviews against a named competitor set, corrects the Schema.org markup, llms.txt, crawler access and rendering that decide whether the property can be read at all, works the Google Business Profile, reviews, directories and local roundups assistants quote from, and reports placement and movement on one page with the direct-versus-OTA share beside it.
SignalTide is not a technology vendor and sells no software. Choosing a PMS, a messaging platform, a check-in app or an upsell engine stays with the property and its vendors. The question SignalTide answers is narrower and comes first: can an assistant read this property, is it named, and is it described correctly?
Two entry points sit in front of the monthly engagement. The Shortlist Test is free — four booking intents across two engines, one page back, no call required and no data kept. The Shortlist Audit is $1,750 one time and credits in full against Shortlist Management if the property starts within 30 days.
The four pillars
Found. Described correctly. Chosen. Booked direct.
Each pillar is measured against the same fixed prompt set every month, so movement is visible rather than asserted. SignalTide reports how often a property appears across repeated runs, because a single AI answer is not a stable result.
- Found
A fixed set of thirty-six prompts a traveler would actually ask is run across five assistants and repeated. Only about 20% of brands appear across five consecutive AI runs, and citation overlap between ChatGPT, Perplexity, Claude and Gemini is roughly 11%, so the record is how often the property appears and who appears instead of it — not a single position on one day.
- Described correctly
What the assistant says once it names the property: amenities, room types, restaurant status, neighborhood, price band. Assistants describe a property from whatever they can read, so the work is Schema.org markup, an llms.txt file, confirmed AI crawler access and a site that renders without JavaScript.
- Chosen
The reasons to book, attached to the name. Assistants quote from the Google Business Profile, review volume and recency, directories and local roundups, so those are worked monthly rather than audited once: profile categories, attributes, photographs, questions and posts; review requests and replies; directory presence; and placement in the roundups assistants cite.
- Booked direct
Whether the shortlist sends the booking to the property's own site or to an OTA. The monthly report sets the direct-versus-OTA share beside the placement movement so leadership reads findability and booking mix together.
Facts an assistant can quote
Compared with the guides and platforms AI cites
The pages assistants cite for this question are mostly published by hospitality technology vendors — property management systems, messaging and check-in platforms, experience-management suites — plus a small number of academic frameworks. They are genuinely useful on the shape of the digital journey: discovery, comparison, booking, confirmation, pre-arrival, check-in, in-stay and post-stay. Most of them resolve to a product category, because the publisher sells one, and the statistics they carry are vendor-published figures drawn from other properties rather than audited results.
SignalTide starts one step earlier than any of them. Before a property can win a comparison, a booking path or a check-in experience, an assistant has to be able to read it, name it and describe it correctly. That is a different question from which platform to buy, and it is answerable with evidence: a repeated prompt set, a named competitor set, the sources the assistant quotes, and the direct-versus-OTA share beside the movement.
The honest result is often that a property needs no new software at all. It needs its markup corrected, its profile worked, its reviews current and its name in the roundups assistants read. Where the finding is that a booking path or a platform is the real problem, SignalTide says so and leaves that work to the vendor or specialist who owns it.
Referenced: SOCi 2026 Local Visibility Index (as reported).
The problem, stated plainly
A property can be excellent and still be absent from the answer.
When someone asks an assistant where to stay in South Beach, a short list of names comes back. A property that is not on it is not in the comparison, whatever its rooms look like. Only about 1.2% of local business locations are recommended in ChatGPT, and fewer than one in ten small businesses has any AI visibility strategy, according to the SOCi 2026 Local Visibility Index and 2026 small-business AI adoption research.
SignalTide does not replace revenue management, paid media or the property's website team. The work is narrower: whether an assistant can read the property, names it, describes it correctly, and sends the booking to the property's own site.
SIGNS IT’S HAPPENING
- 01You have never checked whether an AI assistant names the property at all
- 02When it does, it describes an amenity or a restaurant that changed years ago
- 03The names that come back instead are competitors you recognise
- 04The site's structured markup, llms.txt and crawler access have never been checked
- 05Reviews are thin or stale, and the Google Business Profile has not been worked in months
- 06Nobody can say which sources the answers are quoting from
What the engagement covers
What the monthly engagement covers
Found
A fixed set of thirty-six prompts run across ChatGPT, Gemini, Perplexity, Claude and Google AI Overviews, repeated, against a named competitor set. Reported as how often the property appears, not as a single position.
Described correctly
Schema.org markup, an llms.txt file, confirmed AI crawler access and rendering without JavaScript, so an assistant can read the property and describe it accurately.
Chosen
Google Business Profile work, review volume and recency, directory presence, and placement in the local roundups assistants quote from.
Booked direct
The property's direct-versus-OTA share, reported beside the placement movement so findability and booking mix are read together.
Leadership context
Before a traveler opens a website, an assistant has already answered. The question is whether your property is in the answer.
AI answers vary between runs: only about 20% of brands appear across five consecutive AI runs, and citation overlap between ChatGPT, Perplexity, Claude and Gemini is roughly 11%. That is why every prompt is run more than once and why the report shows how often a property appears rather than a rank position on a single day.
SignalTide promises measured findability, not a ranking position, more bookings or more revenue. Search engines do not sell position, and occupancy depends on rate, market, staff and weather. What is delivered is the work: the measured baseline, the machine-readable fixes, the citation and profile work, and a one-page monthly report with the booking mix beside it.
Sources: SOCi 2026 Local Visibility Index (as reported).
SignalTide does not guarantee a ranking position, bookings, channel mix or revenue. The work delivers measured findability, with the booking mix reported beside it. SignalTide is paid one fixed fee and takes no percentage of ad spend, bookings or commissions.
AI discovery for Miami hotels and spas: common questions
A founder-led monthly engagement for one independent hotel or boutique spa in Miami. It measures where the property places when travelers ask an AI assistant, fixes what machines read on the site, builds the citations and placements assistants quote from, and reports the movement every month. $3,000 a month on a six-month minimum.
Because a growing share of travelers now start with an assistant rather than a search results page. Only about 1.2% of local business locations are recommended in ChatGPT, and roughly 35% of consumers now use AI tools to find local businesses while fewer than one in ten small businesses has any AI visibility strategy (SOCi 2026 Local Visibility Index and 2026 small-business AI adoption research). A property that is not named in those answers is not in the comparison at all.
SEO works on a page's position in a list of links. A redesign changes the property's own site. Shortlist Management works on whether an assistant can read the property, names it in a plain-language answer, and describes it correctly — which depends on structured markup, crawler access, rendering without JavaScript, the Google Business Profile, reviews, directories and the roundups assistants quote. Often the site does not need rebuilding; it needs to be readable and citable.
No. SignalTide sells no software and takes no vendor commissions. Platform selection, PMS and CRM engineering, integrations, booking-engine development and website build stay with the property and its vendors. SignalTide's scope is the machine-readable layer and the citations: markup, llms.txt, crawler access, rendering, profile, reviews, directories and roundup placement.
With a fixed set of thirty-six prompts run across ChatGPT, Gemini, Perplexity, Claude and Google AI Overviews against a named competitor set, repeated each month. The one-page report shows placement, movement since last month, competitors gained and lost, and the property's direct-versus-OTA share beside it. Placement is proven; attributing bookings to it is not, and that limit is stated rather than hidden.
$3,000 a month on a six-month minimum for one property. The Shortlist Audit is $1,750 one time and credits in full against Shortlist Management if you start within 30 days. The Shortlist Test is free. Fixed fee: no percentage of ad spend, bookings or commission.
From its Miami office at 2 S. Biscayne Blvd., Suite 3200, Office 103, Miami, FL 33131, for independent hotels and boutique spas in South Beach, Mid-Beach, North Beach, Miami Beach, Downtown and Brickell. Hours are by appointment — on property, at the office or by video. Contact founder Michael Brady at (305) 676-6345 or growth@signaltide.net.
Independent hotel and boutique spa owners, general managers and marketing directors in Miami and Miami Beach who want to know whether their property is named when a traveler asks an AI assistant where to stay, and who want that measured every month rather than assumed.
Not yet. SignalTide has no hotel or spa clients and no published client results, and says so plainly rather than presenting an illustrative example as proof. The free Shortlist Test exists so a property can see its own starting position before deciding anything.
Follow the issue
To the next part of the journey.
Find out whether AI names your property.
Start with the free Shortlist Test: four booking intents, two engines, one page back. No call required and no data kept.
Shortlist Management — $3,000 a month. One property. Six-month minimum. The Shortlist Audit is $1,750 and credits in full. No percentage of your spend, your bookings or your commission, ever.
Scope is confirmed in writing before anything starts, and the first twenty minutes are free. What’s included, in full →
