- SCORE
- HISTORY
- LABELS
- HUMAN
Short answer
How can I check a user's fraud risk with my AI?
Your agent pulls live fraud scores and decision histories directly from Sift to build a risk profile. You get a clear view of why a user was flagged or cleared. This breakdown shows you exactly how to pull that data.
Risk assessment outcomes.
Where risk data lands.
Your agent turns raw Sift data into actionable intelligence.
SCORE
Real-time risk scores
The AI fetches current fraud scores to show how risky a specific user is right now.
HISTORY
Decision audit trails
Your agent pulls past decision logs so you can see the reasoning behind previous fraud flags.
LABELS
Fraud classification
The AI identifies specific fraud labels assigned to a user to categorize their behavior.
HUMAN
Manual review
The AI flags high-risk patterns that require you to step in and make the final call.
The workflow
What your AI does when a risk check is requested.
The agent acts as your investigator, pulling data from Sift to answer your questions.
Fetch risk scores
The agent requests the current fraud score for a specific user ID.
get_user_fraud_scoreReview past decisions
The agent pulls a list of previous decisions made on the account to find patterns.
list_user_decision_historyIdentify fraud types
The AI looks up specific fraud labels to see if the user matches known bad actors.
get_user_fraud_labelsSummarize risk
The agent combines the score, history, and labels into a single risk summary for you.
get_user_fraud_scorelist_user_decision_history
Try it
Copy these to start.
Use these prompts to begin your investigation.
Starting points
Replace the bracketed text with your actual user IDs or account details.
Sift Connector
You're all set. Choose your MCP client and follow the setup instructions.
https://edge.vinkius.com/vk_preview_X6jrj9G3NJ6qSu4scGacFQiORtOckhn98ppSzoMs/mcpClaude Desktop
Follow the steps below to connect in seconds.
- 1In Claude Desktop, open Settings → Connectors.
- 2Click “Add custom connector” and paste the connector link above as the remote MCP server URL.
- 3Click Add and start a new chat — Sift capabilities are ready to use.
{
"mcpServers": {
"sift-chargeback-mcp": {
"url": "https://edge.vinkius.com/vk_preview_X6jrj9G3NJ6qSu4scGacFQiORtOckhn98ppSzoMs/mcp"
}
}
}Get the current fraud score for user ID [ID_NUMBER].
Show me the decision history for user [USER_ID] to see why they were flagged.
What are the fraud labels currently associated with user [USER_ID]?
Summarize the risk profile for user [USER_ID] based on their score and recent decisions.
Claude
ChatGPT
Cursor
VS Code
Windsurf
Claude Code
JetBrains
Cline
Start here
Connect Sift once, then ask.
One connection links Sift to your agent. Your credentials stay encrypted, and you can start asking questions immediately.
Connect Sift to your AIFAQ
How the investigation behaves
- 01
Can my AI change a fraud decision in Sift?
No. The AI can only read scores, histories, and labels. You must log in to Sift to manually change a decision.
- 02
Does the AI automatically block users?
No. The AI provides the data and the risk profile, but it does not take enforcement actions like blocking accounts.
- 03
Can I see the exact reason a score was generated?
Yes, by pulling the decision history and fraud labels, the AI can explain the context behind a score.
- 04
How often is the fraud score updated?
The AI pulls real-time data from Sift, so you see the most current score available in the system.
- 05
Can the AI delete user fraud records?
No. The tools only allow for reading data. The AI cannot delete or modify any records in Sift.
More questions about Sift? See everything the Sift Connector can do
Connect Sift to Claude, Cursor, ChatGPT & more
