Understanding the Potential claim detector agent

The Potential claim detector agent scans unread guest conversations in your Guesty Inbox for direct reports of physical property damage or harm to a person, and flags them for your review.

The agent is read-only. It surfaces potential claims for your review but doesn't determine insurance coverage or eligibility. Learn more about Guesty Shield and managing a Guesty Damage Protection claim.

Understand the Potential claim detector agent

During each run, the agent does the following:

1. Scans unread guest inbox conversations and reviews eligible guest messages from the configured 24-hour window.
2. Detects direct reports of physical property damage or harm to a person.
3. Flags each conversation that contains a potential claim for your review.
4. Provides a scan result in the activity feed. A "no-claims" result is only shown when all eligible conversations were classified. If conversation coverage or classification is incomplete, the report states that the scan was incomplete instead of showing a no-claims result.

Run the Potential claim detector agent manually

In addition to the scheduled run, you can trigger the agent manually at any time from the Agents hub.

Step by step:

1. Sign in to your Guesty account.
2. In the side navigation menu, select Agents hub.
3. At the top, select the Operations tab.
4. Click the Potential claim detector tile.
5. Under "When it runs", click Run now.

Review the results and status of the run in the activity feed.

Review the activity feed

Follow the steps below to review your activity feed.

Step by step:

1. Sign in to your Guesty account.
2. In the side navigation menu, select Agents hub.
3. In the left submenu, click Agent activity.
4. Click the relevant row with the agent's name to review the report details in the right-side panel.

Understand the report status and details

The report will have one of the following statuses:

• Done: The agent classified all eligible conversations and found no potential claims.
• Action needed: The agent flagged one or more potential claims for your review.

When a potential claim is found, the report shows each flagged conversation with the following details:

• Exact evidence from the conversation
• Related reservation context
• Detected claim category
• Urgency level