Venue Extension

The hospitality extension for describing physical locations where bookings take place.


What is Venue?

A Venue represents any physical establishment that offers Bookables—hotels, restaurants, bars, spas, activity centres. It's the container that holds inventory and provides the context for what guests can expect.

Venue is a namespace extension to A2A Agent Cards that defines what AI agents need hospitality venues to express in order to make defensible recommendations and complete bookings.

Venue does not define a new protocol. It defines a structured vocabulary within existing A2A Agent Cards, extending the base Bookable pattern with hospitality-specific concepts.


Core Concepts

Venue extends the Bookable base pattern with additional blocks specific to physical hospitality establishments.

Identity

Verifiable existence: name, location, coordinates, domain ownership, and company registration.

Vibe

Subjective character structured for matching: essence, atmosphere, energy, formality.

Attributes

Factual, quantified characteristics: rooms, facilities, accessibility, policies.

Evidence

Proof supporting claims with provenance, verification status, and confidence scores.

Fit

Explicit intent alignment—what the venue is good for, and critically, what it is not.

Units

Bookable sub-entities: rooms, tables, areas. Each inherits from venue level.

Neighbourhood

Location context beyond coordinates: setting, walkability, proximity to points of interest.

Presentation

A2UI components for human-in-the-loop confirmation when agents present options.


Type Hierarchy

Venue defines a polymorphic type system enabling agents to work at different levels of abstraction.

Bookable (base)
└── Venue (hospitality)
    ├── Accommodation (Hotel, BnB, Hostel)
    ├── Eat (Restaurant, Cafe, Bar)
    ├── Experience (Museum, Tour, Activity)
    └── Service (Spa, Salon)

All types inherit from Bookable. Agents MAY interact with any Bookable without knowing its specific type—if it's Bookable, you can book it.

Note: Stay is the booking lifecycle spec, not a venue type. When you book an Accommodation venue, the booking record follows the Stay lifecycle.


How Agents Use Venues

When planning a trip, agents query Venue data to make defensible recommendations.

Match Intent to Fit

Query fit.strong and fit.weak to find venues that align with user intent. A "romantic weekend" request should match venues declaring that intent as strong fit.

Verify Evidence Quality

Check evidence blocks for convergence, confidence, and verification status. Prefer claims with DMO verification or human observation over unverified self-reports.

Understand the Vibe

Use vibe to match subjective preferences. Is the guest looking for "lively" or "peaceful"? "Formal" or "casual"?

Check Facilities

Query attributes for specific requirements: pet-friendly, EV charging, accessible rooms, check-in times.

Present with Confidence

Use presentation components to render rich UI for human confirmation. Quote from answers to explain why this venue was recommended.


Venue Conformance Levels

Not every implementation needs everything. Venue defines progressive conformance levels.

LevelRequired BlocksUse Case
Venue Coreidentity, evidence, fit, actionsMinimum viable implementation
Venue ExtendedCore + vibe, attributes, units, neighbourhoodRich discovery
Venue CompleteExtended + presentation, answersFull agent experience

Start with Venue Core. The additional blocks accelerate agent decision-making but aren't required for basic functionality.


Why Declare Weak Fit?

Most hospitality systems only describe what venues are good for. Venue requires declaring what venues are NOT good for.

This matters because:

  • Prevents bad recommendations before they happen
  • Reduces book-then-complain-then-refund friction
  • Signals honesty, which builds trust
  • Improves agent recommendation precision

A venue declaring fit.weak: nightlife signals to agents not to recommend it for hen parties—before anyone books and is disappointed.


Learn More

Specification

Full technical specification with all field definitions and requirements.

Examples

Working JSON examples and complete venue payloads.