Finnosis Fravix: stop fraud before money leaves the bank
Real-time fraud and AML management for every channel. Twelve configurable rules and a machine-learning model score each transaction as it happens, and investigators get the alerts, cases, link analysis and reports to act on it.
Watch Fravix decide
Each transaction goes through four steps in milliseconds: features, rules, machine-learning score, decision.
1 · Transaction
2 · Rules engine
3 · Risk score
4 · Decision
Simulation in your browser with sample transactions. Rule names, severities and the ALLOW / CHALLENGE / HOLD / BLOCK bands follow the Fravix engine; the figures are not customer data.
Rules and machine learning, fused into one score
Eleven signals per transaction
Amount against the account baseline, velocity over one hour and a day, distance from the last location, new device, new payee, merchant risk, time of day, dormancy, cross-border and channel risk.
Twelve configurable rules
Velocity, impossible travel, large amount, deviation from baseline, structuring, new device, new payee, dormant account, off-hours, watchlist, high-risk merchant and high-risk corridor.
A trained risk model
A logistic-regression model scores the fraud probability; an anomaly score catches unusual combinations the rules do not name.
Allow, challenge, hold or block
Rules and model combine into a 0–100 risk score. A critical watchlist match always blocks. Each decision carries the reasons behind it.
What happens to the transaction
| Decision | Risk score | What Fravix does |
|---|---|---|
| ALLOW | below 40 | Approved and posted |
| CHALLENGE | 40 – 64 | Step-up authentication: OTP or 3-D Secure |
| HOLD | 65 – 84 | Held for review by a fraud analyst, alert raised |
| BLOCK | 85 and above, or any critical rule | Blocked in real time before posting, alert raised; three blocks in 24 hours freeze the account and its cards and open a case |
Real screens from the application
Screens captured from the running application with simulated channel traffic.
Did you just pay USD 780.00 at a crypto merchant? Tap to confirm.
We paused this payment because it doesn't look like your usual spending.
Touch the fingerprint sensor,
or enter the OTP we sent you
A new card is on its way
The customer is part of the defence
When Fravix challenges or holds a payment, the customer is asked straight away on the channel they use, and can confirm, block their card or report fraud themselves.
Push notification, SMS or WhatsApp message with Yes / No.
OTP, 3-D Secure or biometric before a challenged payment goes through.
Freeze or unfreeze a card in the app, on USSD or in a chat.
Report a transaction and get a case reference, without calling.
Staff see the alert and case status when a customer calls or walks in.
Real screens for each team
Captured from the running application. Each role signs in to its own menu; permissions are enforced on the server, not just hidden.
Front-end and back-end users
| User | Front or back end | What they do in Fravix |
|---|---|---|
| Customer | Front end | Confirms or rejects flagged payments, completes step-up, freezes a card, reports fraud. |
| Branch, agent and contact-centre staff | Front line | Verify the customer, raise a fraud report, see alert and case status. |
| Fraud analyst | Back end | Works the alert queue, investigates, uses link analysis, opens and updates cases. |
| Risk manager | Back end | Watches the dashboard and trends, tunes rules and thresholds, manages watchlists. |
| Compliance officer | Back end | Reviews cases, confirms fraud or false positives, reviews and files suspicious activity reports. |
| Administrator | Back end | Manages users, roles and settings; full audit trail. |
| Viewer or auditor | Back end | Read-only dashboards and reports. |
Configure detection without a release
Rules, thresholds, decision bands, automatic actions, lists and roles: set by the bank, recorded in the audit trail.
Rules & thresholds Changes apply live and are audited
Decision bands Risk score 0–100
A critical rule (watchlist match) always blocks.
| Automatic action | Setting |
|---|---|
| Auto-freeze account and cards | 3 blocks within 24 hours |
| Open a case on freeze | Yes · priority Critical |
| Step-up method for CHALLENGE | OTP or 3-D Secure |
| Alert on HOLD | Fraud analyst queue |
Watchlists & blacklists Matched on every transaction
| Type | Value | Reason | Severity |
|---|---|---|---|
| Device | dev-blk1a2b3c | Linked to account-takeover cases | HIGH |
| IP address | 185.220.••.47 | Known anonymiser exit node | HIGH |
| Beneficiary | AC9000••22 | Mule account flagged by partner bank | HIGH |
| Customer | Screening hit | Sanctions fuzzy match under review | MEDIUM |
Roles & access Each role sees only its screens
| Permission | Admin | Risk mgr | Analyst | Compliance | Viewer |
|---|---|---|---|---|---|
| Dashboard | ✓ | ✓ | ✓ | ✓ | ✓ |
| Work alerts and cases | ✓ | ✓ | ✓ | ✓ | — |
| Generate SAR drafts | ✓ | ✓ | — | ✓ | — |
| Edit rules and thresholds | ✓ | ✓ | — | — | — |
| Manage watchlists | ✓ | ✓ | — | ✓ | — |
| Audit trail | ✓ | ✓ | — | ✓ | — |
Front end and back office, together
The customer confirms a suspicious payment on their phone while the back office configures the rules that caught it.
Anti-money laundering, from onboarding to the regulator
Fravix covers the whole AML process: know your customer at onboarding, keep screening, monitor every transaction, investigate, report and keep the evidence.
Customer due diligence
Identity verification at onboarding, customer risk rating and enhanced due diligence for high-risk customers.
Sanctions and PEP
Customers and payments screened against sanctions and politically exposed person lists, with fuzzy matching and recorded decisions.
AML scenarios
Structuring, dormant-account reactivation, high-risk corridors, rapid movement of funds and deviation from the customer's profile.
Investigation
Alerts grouped into cases, with notes, link analysis and a confirmed or false-positive outcome.
SAR / STR
Suspicious activity report drafted from the case for the compliance officer to review and file with the financial intelligence unit.
Audit and retention
Append-only audit trail of every decision and action, kept for the regulator's retention period.
Find the mule ring behind the transaction
Fraud rarely comes from one account. Fravix links accounts that share a device, an IP address or a beneficiary, so investigators see the network, not just the alert.
- Time windows of 24 hours, 72 hours or 7 days
- Hubs highlighted where many accounts meet
- Open any account to see its devices and history
From the operations wall to the audit trail
Dashboard, alert queue, cases, link analysis, rules console and audit trail, screen by screen.
Everything a fraud and AML team needs
Six channels
Card / POS, ATM, internet banking, mobile banking and UPI, wire transfer and SWIFT, merchant payment gateway.
Auto-freeze
Three blocked transactions on an account within 24 hours freeze the account and its cards and open a case automatically.
Case management
Investigation workflow from open to confirmed fraud or false positive, with notes, assignment and priority.
SAR drafts
A suspicious activity report drafted from the case for the compliance officer to review and file.
Watchlist & blacklist
Devices, IP addresses, beneficiaries, accounts and customers under scrutiny, matched on every transaction.
Trend forecasting
Flagged-transaction trend with a short-term forecast, overall and by channel, to spot an attack building.
Roles
Administrator, risk manager, fraud analyst, compliance officer and viewer, each with its own permissions.
Audit trail
Every engine decision and every analyst action, append-only, ready for internal audit and regulators.
Feedback loop
Each case an analyst confirms or dismisses labels the transaction, building the data for the next model retrain.
See Fravix on your channels
A live demo with simulated traffic shaped like your bank's channels and fraud patterns.