What is an Echo?

An Echo is how AI agents understand a venue. It's structured information about who a place is, what it offers, and how to book it.

Why structured data matters

When someone asks an AI agent to "find somewhere romantic for an anniversary," the agent needs to understand what "romantic" actually means for different venues. A candlelit bistro is romantic. A sports bar isn't. But that distinction isn't obvious from a website or a star rating.

Structured data gives AI agents the vocabulary to understand these differences. Instead of parsing marketing copy and guessing, the agent reads clear, organised information: this venue has an intimate atmosphere, suits couples, works well for special occasions.

The result: better matches. Guests who actually fit what the venue offers.

An example

A boutique hotel might have this in its Echo:

• Character: intimate, design-forward, locally-rooted
• Suits: couples, creative retreats, digital detox
• Occasions: anniversaries, milestone birthdays, reconnection trips
• Evidence: verified by regional tourism board, consistent across multiple data sources

When an agent searches for "somewhere quiet to disconnect for a few days," this hotel surfaces. A chain hotel with conference facilities doesn't.

The four pillars

An Echo is built on four pillars. Each covers a different aspect of what AI agents need.

The Groundry

Groundry is Selfe's service that builds Echos. It gathers information from multiple sources, structures it to spec, and keeps it current as things change.

Venues don't write their Echo manually. Groundry does the work, drawing on existing data, reviews, and direct input where needed.

Open standards

Echo implements the Agentic Booking specifications. The specs are open—anyone can build compatible infrastructure.

Selfe is an implementation of these specs, not the only possible one. Our role is to make it easy: we build Echos, maintain the catalog, and handle the complexity so venues and agents don't have to. Your venue data follows the open spec—export it anytime, and another implementation could use it.

This matters because:

• No lock-in — Your data follows an open standard
• Ecosystem growth — More implementations mean more agents, more reach
• Trust — Open specs can be audited and improved publicly