Detect Fraudulent Receipts Agent is designed to identify possible receipt fraud from customer attachments. It helps teams flag suspicious receipts for review using their own fraud indicators.
Before you start
Make sure your Help Desk account has:
- Access to Autopilot.
- Ticket categories where this AI Agent should run.
- A clear rollout plan for whether your team should review the AI Agent's work first.
Hire the AI Agent
- Open Autopilot > Home.
- Find Detect Fraudulent Receipts Agent.
- Select Hire Agent.
The setup flow opens on the Onboard step.
Configure the AI Agent
- Give your AI Agent a name.
- Enter a description that explains what the AI Agent should do.
- Complete the remaining setup fields.
Complete the setup fields
- Ticket categories where the AI Agent runs: Choose the ticket categories the AI Agent should handle. Start narrow, then expand after reviewing early activity.
- Supervised Mode or Autopilot Mode: Choose Supervised Mode when support agents should review early work. Choose Autopilot Mode only after your team has reviewed the AI Agent's results and confirmed its setup.
- Tag to apply when potential fraud is detected: Choose a tag that makes the AI Agent's output easy to filter, report on, or route later.
- Custom instructions: fraud indicators the AI Agent should look for: Add policy rules, examples, exceptions, and tone guidance that help the AI Agent make the right decision for matching tickets.
If a field you expect is missing, check the related Help Desk setup before continuing.
When to use this receipt agent
Use Detect Fraudulent Receipts Agent for provider-neutral receipt review. It applies the fraud indicators you enter during Train & Test without adding built-in guidance for a particular receipt provider.
For Amazon receipts, use Amazon receipt fraud detection agent, which includes additional built-in Amazon-specific guidance.
Review before deployment
- Select Continue.
- Review the Deploy step.
- Confirm the AI Agent name, description, ticket categories, operating mode, and any additional setup fields.
- Select Deploy Agent only after you confirm the configuration is correct.
What happens after deployment
The AI Agent appears in Autopilot > Hired Agents. When matching work reaches the selected scope, the AI Agent uses its configured settings to complete the intended action.
After the AI Agent processes matching work, open Agent Activity Log to confirm its status and review the result. If the result needs adjustment, open Hired Agents and update its configuration.
Troubleshooting
- The AI Agent does not run on the expected tickets: Review the selected ticket categories and operating mode.
- The AI Agent output is incomplete: Review the setup fields and add clearer policy, tag, status, or custom instruction details.
- The AI Agent should not act automatically yet: Use Supervised Mode when available while your team reviews behavior.