Dashboard Last 30 Days
Hourly Volume 7 days
Recent Critical Alerts
Fraud Alerts Live · Last 24 Hours
1st Payment Alerts Newly Activated Merchants — First PSP Transaction
#first-time-merchant-payment-watchlist
● Connected via Claude
Velocity Monitor Rolling 5-Min Windows
Case Queue 🏠 Floorzap · Fraud/Risk Only
Geographic Conflicts Last 30 Days
FI / Billing / IP Conflicts
Repeat Shoppers — Multi-State
Behavioral Anomalies Last 30 Days
Copy-Paste & Keystroke Patterns
Rapid Interval Pairs (<5 min)
Fraud Networks Connected Fingerprint Clusters
Brain Adaptive Learning Engine
Merchant Age — Amount Thresholds
Based on Onboarded At (Col H) · Amounts above limit trigger fraud rule · ≥$35,000 always flags Business RiskRule Weights — Live Learning
Verify → weight ↓ | Flag / Confirm → weight ↑ | Range: 0.5× – 2.0×Merchant Timezones
Auto-detected from state. Edit if incorrect.HubSpot Integration
ConnectedCreates HubSpot tickets on every Open Case / Confirm Fraud action and syncs the Case Queue in real time via your HubSpot Private App token. A local proxy fallback is also supported.
▸ Advanced — override portal / pipeline / stage IDs
📊 Weekly Fraud Intelligence Report Confirmed fraud patterns · predictive analysis · weight recommendations
Runs every 7 days — analyzes all confirmed fraud verdicts, clusters common rule co-occurrence patterns,
surfaces dormancy and timing signals, and outputs Bayesian weight recommendations your PostgreSQL
scoring engine can apply immediately. Open the exported .txt file in any text editor, or run:
psql -d floorzap_fraud -f weekly_fraud_intelligence_*.txt
💾 Backup & Restore Action history feeds the SQL scoring engine
Every fraud verification, confirmation, dismissal, and refinement is logged here and exported as two files:
a JSON snapshot (full IndexedDB state) and a SQL feedback file
with INSERT statements ready to apply directly to your floorzap_fraud PostgreSQL database.
The SQL file calls SELECT refresh_ml_features(); automatically so the scoring engine learns immediately.
Every verified, flagged, and confirmed-fraud action is saved to a separate
localStorage journal that persists even when IndexedDB is cleared. On every page load and after every import, verdicts are automatically re-applied to your transaction data — so re-opening this file never erases your reviews.
Import Data
Account column — FloorZap rows are deduplicated automatically. Accepts multiple files.
amount · transaction_at · shopper_email or emailRecommended:
store_id · card_type · card_number_last4 · cardholder_name · billing_state · fi_state
🧠 Import Confirmed Fraud Training Data
Brain LearningUpload a CSV or Excel file of historically confirmed fraud transactions in Adyen export format. The brain will import each case, score it against all rules, then boost the weight of every rule that fired — permanently improving detection of similar patterns in future transactions. Chargeback cases automatically update shopper and card risk history.
Platform Merchant Rosters
Floorzap · 353 AccountsThe Floorzap merchant roster is pre-loaded. Apply it to create merchant stubs — they will auto-link when you import transaction data.
⚖️ Import Historical Chargeback Data
Dispute Cortex
Import historical platform chargeback records to populate the Dispute Cortex pipeline with past cases.
ACH and credit card chargebacks are tracked separately for accurate reporting.
Required columns: merchant_name · amount · network (visa/mc/amex) · reason_code · dispute_date · status (won/lost/new)
Optional: txn_id · notes · payment_type (ach/cc — used to separate ACH vs card chargebacks)
🔄 Export & Restore Brain Data
Migration · BackupClick below to save all learned weights, verdicts, and dispute data to a file. Do this while the dashboard is open in your browser.
Open the Transaction IQ app, come back to this same Import tab, and click below to load the file you just exported.
Export Scored Transactions
Download all transactions as a CSV with the Teamfront Score column filled in — ready to upload back to Adyen or share with your team.
📡 Live Feed — Adyen Webhooks
Coming SoonWhen your Adyen webhooks are configured, run this Python script on your Mac (Python 3 is pre-installed). It will receive webhooks and feed them to the brain automatically — no other installation needed.
Save this as webhook_receiver.py on your Desktop:
Point your Adyen webhook notification URL to: http://localhost:4747/webhook
CSV Column Reference — Adyen Export Format
| Adyen Column Header | What the Brain Uses It For | Required? |
|---|---|---|
Shopper Email | Shopper identity key | ✅ Required |
Value | Transaction amount | ✅ Required |
Creation Date | Transaction timestamp | ✅ Required |
Store Description | Merchant identifier | ⭐ Recommended |
Shopper Name | Shopper full name | ⭐ Recommended |
Account Holder Name | Name on card (name-match rule) | ⭐ Recommended |
Metadata | Extracts CustomerName JSON field for name-match | ⭐ Recommended |
Payment Method | Card type — auto-normalized (visa→credit, ACH→ach, etc.) | ⭐ Recommended |
Shopper Billing Address | Billing address — state parsed automatically | ⭐ Recommended |
Shopper Country | IP-derived country code — used as Shopper Country (DO = auto-blocked at score 100) | ⭐ Recommended |
Issuer | Financial institution name | ⭐ Recommended |
TimeZone | Merchant timezone for after-hours rule (EST/CST → IANA) | ⭐ Recommended |
Shopper IP | IP address / country mismatch detection | Optional |
Account | Account identifier | Optional |
PSP Reference | Unique transaction ID (deduplication) | Optional |
Currency | Currency code (USD, EUR…) | Optional |
Status | authorized / declined / captured… | Optional |
Risk Score | Adyen's own risk score (stored for reference) | Optional |
Teamfront Score | ⬅ OUTPUT — brain writes its fraud score here on export | Output only |
Portfolio Overview
Chargeback Loss by Product
Fraud Flags by Month
https://floorzap.com/adyen/webhook → FraudGuard scorer