Your hotel needs a
Canonical AI Profile.

A hotel already has a website for guests, booking systems for transactions, and a wide digital footprint distributed across OTAs, directories, maps, reviews, corporate pages, historic content, and third-party descriptions. What it usually does not have is one governed operating model that defines, in machine-usable form, what the property actually is, what it can support, where its capabilities stop, what evidence supports material claims, and where stable hotel truth must hand the customer to live commercial systems.

The Canonical AI Profile provides that operating memory. It sits at the center of the hotel's Governed AI Identity. Evidentity uses it to structure Operational Truth, Scenario Architecture, evidence, policies, restrictions, Claim Authority, boundaries, commercial routes, and knowledge states before those elements are projected into first-party AI-facing surfaces and tested against real recommendation behaviour.

The Profile is not another description of the hotel. It is the canonical model from which Recommendation Infrastructure operates.

Without a Canonical AI Profile, the property remains distributed across many representations created for different audiences and maintained by different systems. The hotel website may describe the experience well, the booking engine may hold current inventory, an OTA may carry an older policy, a directory may omit important operational detail, and internal teams may know capabilities that have never been expressed clearly in public machine-facing form.

AI systems must then reconstruct a usable version of the hotel from that environment. Recognition alone is insufficient. When a traveller asks whether the property can support a late arrival, a particular family configuration, accessible accommodation, secure parking, remote work, a residential training programme, or another constrained requirement, the system must determine not merely whether the hotel exists but whether it remains a credible candidate for that exact situation.

This is where otherwise strong hotels can lose Recommendation Participation. The physical capability may exist while the recommendation environment cannot establish it clearly enough, or a competing property may simply present a stronger combination of Scenario Fit, evidence, operational legibility, and commercial clarity. The problem is therefore not simply visibility. It is the gap between operational reality and decision-grade representation.

The unmanaged web presents the property as a collection of signals. AI systems must decide which descriptions belong together, which information is current, which claims are official, how broad statements should be interpreted, and where exceptions apply. This reconstruction can remain workable for simple requests and become increasingly fragile as the user adds operational constraints.

The Canonical AI Profile changes the business-facing side of that problem. It establishes a governed model in which identity, Stable Truth, evidence, Scenario Fit, limitations, unknowns, Claim Authority, and commercial handoff are represented deliberately rather than left entirely to inference.

The purpose is not to prevent independent AI systems from interpreting the wider web. They will continue to use their own sources, retrieval systems, and reasoning. The purpose is to give the hotel something it previously lacked: a coherent first-party operating representation against which its wider digital presence can be published, maintained, and tested.

The external system still interprets. The hotel no longer leaves all of its meaning to reconstruction.

The Profile is not a PDF, a database export, a schema file, or a static piece of content. It is the controlled canonical model from which other parts of the Governed AI Identity can remain synchronized.

A mature machine-facing identity requires more than a collection of positive claims. It requires governance over which facts are authoritative, which statements are conditional, what is explicitly unavailable, what remains unresolved, and which live system owns the next stage of the decision.

Evidentity calls this logic the Recommendation Contract. It is not a legal agreement with an AI provider. It is the operating boundary governing the machine-facing representation of the hotel. It defines which information can be represented as Stable Truth, which claims require conditions, which source or authority establishes them, and where the recommendation layer must hand the customer to an authoritative transactional system.

This makes the Profile useful not because it claims to know everything, but because it distinguishes clearly between what is known, what is bounded, what is unknown, and what belongs elsewhere.

The Canonical AI Profile is not created by asking a model to scrape the public web and decide what the property probably means. External information can contribute evidence and reveal conflicts, but the hotel remains the authority for the parts of Operational Truth it legitimately controls.

Evidentity structures that reality, applies claim and evidence governance, connects it to Scenario Architecture, maintains the canonical model, and projects approved information into the broader Recommendation Infrastructure. This gives the property continuity. When the underlying business changes, the canonical state can change with it.

A revised parking condition, new room configuration, changed arrival procedure, new meeting capability, updated accessibility detail, or altered service restriction does not need to become another isolated fact drifting through the digital estate. It enters the governed identity and can propagate through the controlled surfaces Evidentity operates. Currentness is therefore tied to operational change, not cosmetic publishing activity.

One of the strongest properties of the Canonical AI Profile is not merely what it contains, but what it deliberately refuses to treat as canonical. Identity, persistent policies, infrastructure, restrictions, Scenario Boundaries, stable service conditions, operating capabilities, declared absences, evidence, and official handoff routes can usually be governed as Stable Truth. Current room prices, real-time inventory, availability, dynamic packages, and other volatile commercial conditions belong to Live State.

The Profile therefore does not attempt to become a booking engine. It establishes enough reliable context for a recommendation decision and then identifies the authoritative continuation path when live commercial information becomes necessary. This separation protects both accuracy and usability. The hotel retains control over the systems designed to manage inventory and pricing, while the AI-facing identity provides a clean bridge from recommendation to transaction.

The Canonical AI Profile becomes commercially useful when governed information can be projected into a public machine-facing environment without losing its structure, boundaries, evidence relationships, or connection to the hotel itself.

Evidentity does this through the AI-Facing Publication Layer. A hotel can receive a dedicated first-party AI Site, typically deployed through architecture such as ai.yourhotel.com, where the Governed AI Identity can be represented explicitly for intelligent systems. Where appropriate, that architecture can also be reinforced through a machine-facing layer on the primary hotel domain, such as yourhotel.com/llm, together with structured endpoints, canonical relationships, evidence references, metadata, and official commercial routes.

The distinction is important. The Canonical AI Profile is the governed operating memory. The AI Site and associated machine-facing surfaces are the public projection of that memory. Recommendation Intelligence then measures how independent AI systems actually treat the hotel after publication. Publication is not assumed to equal recommendation. It creates an infrastructure state that can be observed and tested.

The value of a Canonical AI Profile does not come from representing more facts for their own sake. It comes from making those facts usable inside the decisions through which traveller demand is allocated.

A meeting room is a fact. A 60-person residential training programme requiring bedrooms, daily meeting space, group dinners, parking, and straightforward arrival logistics is a Scenario Market. Parking is a fact. A late-arriving project team in which many guests are driving creates a different qualification problem. Accessible rooms are facts. A particular accessible traveller request combines those facts with route, room, bathroom, entrance, and operational requirements that determine whether the property actually fits.

Scenario Architecture connects the fact layer to this commercial layer. The Profile therefore helps preserve more of the hotel's real capability as Candidate Compression becomes more demanding. It does not manufacture Scenario Fit; it makes genuine fit more explicit, evidenced, bounded, and available for evaluation.

A physical hotel contains far more commercial capability than a generic property description can express. Rooms, location, meeting infrastructure, parking, dining, wellness, family configurations, accessibility, arrival flexibility, outdoor assets, service patterns, and operational expertise can each open different Recommendation Territories.

The commercial problem appears when those capabilities exist in the asset but do not survive into the recommendation environment. That difference contributes to the hotel's Recommendation Gap: the distance between the Addressable Recommendation Footprint created by the real property and the Observed Recommendation Footprint visible through independent AI testing. The Canonical AI Profile is one of the foundational mechanisms for reducing addressable parts of that gap. It strengthens the governed identity from which evidence, Scenario Architecture, first-party publication, commercial handoff, Recommendation Intelligence, intervention, and Profile Protection can operate as one system.

The objective is not simply to make the hotel more understandable. It is to make more of the commercial capability already embedded in the asset structurally available for the AI-mediated decisions that capability has earned the right to contest.

Where Evidentity uses a public recommendation protocol, its purpose is to make the publication semantics inspectable rather than to treat the underlying architecture as hidden vendor magic. A public protocol can define how authority, claim state, scenario boundaries, evidence relationships, machine-facing projection, and transactional handoff are represented for external technical review.

REP-01 is the public projection contract supporting the Canonical AI Profile and AI-Facing Publication Layer. It is not presented as proof that independent AI systems are required to adopt or honour the protocol.

The Profile provides memory and governance. The AI Site and wider AI-Facing Publication Layer make approved identity publicly accessible. Recommendation Intelligence observes how independent AI systems include, omit, compare, substitute, qualify, recommend, and route the hotel across commercially meaningful scenarios. Recommendation Control uses those observations to diagnose addressable weaknesses, intervene in the infrastructure Evidentity can legitimately govern, and retest comparable decisions. Profile Protection maintains integrity as the hotel, competitors, public information environment, and AI systems change.

Together, these components form AI Recommendation Infrastructure. The Canonical AI Profile sits at the center because every intervention, publication, evidence relationship, scenario model, and update requires one governed understanding of the business behind it. Without that memory, Recommendation Infrastructure eventually fragments. With it, the hotel acquires a durable operating identity from which participation can be measured, strengthened, and protected over time.