Why AI Agents Need Machine-Readable Trust Signals
When a human customer books a plumber, they can read profile reviews, check certifications manually, and ask questions before confirming. When an AI agent books a service on a user's behalf, it cannot do those things at human speed or with human judgement. It needs structured, verifiable signals it can parse in milliseconds.
The Gera Services Passport solves this by producing a machine-readable trust signal for every verified provider. An AI agent checking a Gera provider receives:
- A confirmed identity (the provider is who they say they are)
- A verified tier (the level of checks passed)
- A warrantyEligible flag (whether a booking will generate a warranty-eligible action receipt)
- A link to the claim cap and eligibility rules for the relevant service type
- A cryptographic signature the agent can validate against Gera's public key
The Badge JSON Structure
Every active Gera Passport produces a response at GET /api/verify/{providerId} that looks like this (Elite tier example):
{
"@context": "https://schema.org",
"@type": "Certification",
"name": "Gera Elite",
"certificationStatus": "Active",
"issuedBy": {
"@type": "Organization",
"name": "Gera Services Ltd",
"url": "https://gera.services"
},
"about": {
"@type": "Person",
"name": "{providerName}",
"identifier": "{providerId}"
},
"additionalProperty": [
{
"@type": "PropertyValue",
"name": "tier",
"value": "elite"
},
{
"@type": "PropertyValue",
"name": "warrantyEligible",
"value": "true"
},
{
"@type": "PropertyValue",
"name": "qualificationsVerified",
"value": "true"
},
{
"@type": "PropertyValue",
"name": "backgroundCheckCleared",
"value": "true"
},
{
"@type": "PropertyValue",
"name": "sustainedRating",
"value": "4.8+"
},
{
"@type": "PropertyValue",
"name": "zeroClaimsWindow",
"value": "24 months"
}
]
}The warrantyEligible field is the primary signal for AI booking agents. A value of "true" means the provider's next booking will generate a signed action receipt covered by the Gera Action Warranty limited service guarantee.
The End-to-End AI Agent Verification Flow
AI Agent or Booking System
Looks up provider by Gera ID or username
The agent queries the Gera Passport API endpoint: GET /api/verify/{providerId}. No authentication is required for the public trust status check.
Passport API
Returns a signed JSON-LD badge object
The response includes the provider name, badge tier, warrant eligibility flag, issue date, expiry policy, and a cryptographic signature. A 200 status means active; 404 means no active Passport.
AI Agent
Reads warrantyEligible field
If warrantyEligible: true, the agent knows that bookings with this provider will generate a signed action receipt covered by the Gera Action Warranty limited service guarantee. The agent can surface this to the user before confirming the booking.
Booking Confirmation
Action receipt generated with warranty signal
On booking confirmation, a signed action receipt is created that includes the warrantyEligible flag, the claim cap for the service type, and the claim window duration. The receipt is machine-readable and accessible via the booking API.
Post-Service
Agent can check receipt status and flag potential claims
After service completion, the AI agent can read the action receipt status. If the service outcome is marked incomplete or disputed, the agent can surface the claim option to the user, or (with appropriate authorisation) open a claim on the user's behalf.
Integration with the Gera Action Warranty
The Gera Services Passport and the Gera Action Warranty are designed as a complementary pair. The Passport is the entry gate: it confirms who the provider is and what tier of verification they hold. The Action Warranty is the outcome guarantee: it describes what happens if a verified provider's service fails.
For AI agents, this two-step structure provides a complete trust loop:
- 1Before booking: verify the provider holds a valid Passport and the warrantyEligible flag is true.
- 2At booking confirmation: confirm the action receipt was generated with the correct warranty cap for the service type.
- 3After service: read the action receipt status; surface claim option to the user if the service outcome is disputed.
This loop means an AI agent can act as a genuine trust-and-recovery layer — not just booking a service, but monitoring the outcome and escalating to the user only when intervention is needed.