Skip to content
  • Home
  • Company
  • AWS Marketplace
  • By Mercurius
    • Hawk – Free CTI
    • Cactus – Free EASM
  • Solutions
    • Discover
    • Defend
    • Exposure Management
    • Govern
    • Fraud Prevention
  • Blog
  • Contact
  • EN
    • PT
    • ES
  • Home
  • Company
  • AWS Marketplace
  • By Mercurius
    • Hawk – Free CTI
    • Cactus – Free EASM
  • Solutions
    • Discover
    • Defend
    • Exposure Management
    • Govern
    • Fraud Prevention
  • Blog
  • Contact
  • EN
    • PT
    • ES
  • Pillar 05 · Fraud Prevention

Fraud moves through your customer’s journey. So does our defense.

Layered protection from the fake site before the click to the transaction, built on what you already have and connected to our 24/7 AI SOC.

  • Banks · fintechs · insurers · retail · Brazil, Chile and the US
Request a fraud-risk assessment

or

Message a Specialist on WhatsApp
ATTEMPT #4471 R$ 0 at risk
tracking
attempt in progress
1.90s
TO PAYOUT
0 attempts stopped before loss — this session

COVERAGE

6 layers

JOURNEY

Pre-attack to scam

SIGNALS

Into the AI SOC

ADOPTION

PoC by use case
Pix fraud
Fake call center
Account takeover
Mule accounts
APP fraud
Deepfakes
Phishing
AML
CMF NCG 538
BCB 538/2025
  • The problem we solve

Every fraud tool sees one signal. Fraudsters use all of them.

A modern scam starts with a cloned site or a fake profile, continues with a stolen identity or a hijacked session, and ends with a transfer to a mule account, often while the victim is on the phone with someone posing as the bank.

Liveness checks, device fingerprinting and transaction engines each see one step. When they work in silos, the fraud passes between them. And with the burden of proof shifting to the bank in Chile and Brazil, every loss costs more.

We design the layers as one defense, around how your customers actually move.

  • Definition

What is customer journey fraud prevention?

Customer journey fraud prevention is the practice of protecting each stage a customer goes through, from first contact to payment, with controls that see different signals: brand impersonation, identity, device and session behavior, critical actions, transactions and the human being manipulated.

Its strength comes from correlation: signals from every layer are combined, so a fraud missed in one stage is caught in the next.

Anti-phishing
Liveness
Device intelligence
ATO
Mule detection
APP fraud
FRAML
Side by side

One more tool will not close the gap. A connected defense will.

Switch the view and compare.

  • ViewThe whole journey, from pre-attack to scamOne stage of the journey
  • SignalsCorrelated in our AI SOC and your SIEMSeparate consoles, separate teams
  • AdoptionLayers added on top of what you already haveRip and replace
  • ProofA PoC per use case before any commitmentA vendor promise
  • FrictionExtra checks only on risky actionsThe same friction for every customer
  • EvidenceForensic trail to contest claims and support refundsLogs scattered across tools
  • OutcomeAn executive report of losses avoidedA count of alerts
  • HOW IT WORKS

Two views of your risk. One decision.

What attackers can see and plan from the outside meets what is vulnerable on the inside, and becomes a single priority.

  1. 01

    Pre-attack

    Fake sites, apps and profiles taken down before the click.

    Phishing, impersonation
  2. 02

    Onboarding

    Identity checked with liveness, document and digital signature.

    Synthetic identities, deepfakes
  3. 03

    Session

    Remote access, malware and the fake call center, detected live.

    RAT, active-call scams
  4. 04

    Critical action

    Device change, high-value Pix or credit, verified at the moment.

    Account takeover
  5. 05

    Transaction

    Real-time scoring, mule accounts and money laundering.

    Mules, APP fraud, AML
  6. 06

    Scam

    AI agents in investigation and in the contact center.

    Victim manipulation
Every signal flows into the Mercurius AI SOC· correlated, investigated and reported
  • Pillar 05 · Fraud Prevention

The layers you need, designed around the ones you have.

We start with a diagnosis and add only what closes real gaps.

Where we start

Fraud Journey Assessment & Managed Fraud Operations

A vendor-agnostic diagnosis of your customer journey: where losses concentrate, what each existing tool covers and which layers close the gaps. Then PoCs by restricted use case, integration of fraud signals into the SOC, rule tuning and an executive report of losses avoided.

DiagnosisLayer architecturePoCSOC integration

When fraud losses grow and nobody owns the whole journey

Anti-Phishing & Takedown

Predictive detection of malicious domains, cloned sites and fake profiles, taken down before the campaign reaches your customers.

When your brand is the bait

Digital Identity & Onboarding

Facial biometrics with liveness, document checks and digital signatures with full audit trail.

When onboarding is fully remote

Device & Session Intelligence

Remote access, emulators, overlays and active calls during the transaction, detected in the app and on the web.

When the fraud happens inside a real session

Action Verification & Anti-ATO

A strong verification only on critical actions, with a forensic trail to contest disputed claims.

When device changes and resets are the entry point

Transaction Monitoring & Mule Detection

Real-time scoring for cards, Pix and transfers, mule account detection and AML on the same engine.

When money leaves faster than you can react

Scam Defense with AI Agents

AI agents that support fraud investigation and coach the contact center during live scams. Available in the United States.

When the victim is the one authorizing the payment

  • Who operates it

The Mercurius fraud team, connected to our SOC.

ARCHITECTURE

Fraud architects

Map the journey, design the layers and define how they coexist with your current tools.

VALIDATION

PoC management

Run each PoC on a restricted use case, with success criteria agreed before it starts.

OPERATIONS

Fraud analysts

Tune rules, review cases and produce the report of losses avoided.

RESPONSE

AI SOC 24/7

Correlates fraud signals with security events and acts when an attack is under way.

What you receive

The answer each team needs.

For your fraud and risk teams

Fewer losses with less friction for good customers.

  • Correlated casesSignals from every layer combined into a single case per incident.
  • Rule tuningThresholds adjusted to your customers, so verification appears only where risk is real.
  • Mule account visibilityReceiving accounts monitored on the way in and on the way out.

For your CISO

Fraud treated as part of security, not a separate silo.

  • Journey coverage mapWhich layer covers each stage, and where the gaps are.
  • SOC and SIEM integrationFraud signals correlated with security events in our AI SOC and your SIEM.
  • Pre-attack intelligencePhishing infrastructure and brand abuse detected before campaigns start.

For your CFO and compliance

Numbers and evidence that hold up.

  • Executive report of losses avoidedWhat was stopped, at which stage and what it would have cost.
  • Forensic trailEvidence to contest claims and support refund processes.
  • Regulatory alignmentControls mapped to CMF NCG 538, Ley 21.673 and 21.719 in Chile and BCB 538/2025 in Brazil.
  • One continuous cycle

Every other pillar produces evidence. Govern turns it into proof.

Because we run your pentests, your SOC and your exposure program, the evidence already exists. Nobody has to chase it.

PILLAR 03 · EXPOSURE

We see it coming

Fraud intelligence and brand protection detect campaigns before launch.

 

PILLAR 05 · FRAUD PREVENTION

We stop it in the journey

Six layers protect each stage, from onboarding to payment.

 

PILLAR 02 · DEFEND

We connect the dots

Fraud signals meet security events in our AI SOC, 24/7.

 

  • FAQ

What fraud and security leaders ask.

Do we need to replace the anti-fraud tools we already have?

No. We start by mapping what each existing tool covers, then add only the layers that close real gaps. Every new layer is designed to coexist with your current stack.

How do you stop the fake call center scam?

By combining layers: session intelligence detects signs such as an active call or remote access while the customer is making a transaction, action verification confirms risky operations, and transaction monitoring flags the receiving account. A single control rarely catches it alone.

How do you detect mule accounts in Pix and transfers?

With real-time transaction monitoring that watches receiving accounts on the way in and on the way out, combined with behavior patterns and AML rules on the same engine.

How does this help with CMF NCG 538 and Ley 21.673 in Chile?

NCG 538 requires reinforced customer authentication, and Ley 21.673 lets the issuer contest claims when there is evidence of intent or gross negligence. Strong verification on critical actions and a forensic trail support both, reducing liability and losses.

 

Biometric data is sensitive. How is it handled?

Under laws such as Chile’s Ley 21.719 and Brazil’s LGPD, biometrics are sensitive data. We assess each provider’s storage architecture and data residency, support the impact assessment and define consent flows before anything goes live.

How can we start without a large commitment?

With a Fraud Journey Assessment and a PoC on a restricted use case, such as device changes or high-value transfers, with success criteria agreed in advance. You decide on scale only after seeing results.

Do fraud signals reach our SOC and SIEM?

Yes. Integrating fraud signals into our AI SOC and your SIEM is part of the service, so fraud and security events are investigated together.

Which countries and sectors do you cover?

Brazil, Chile and the United States, with a focus on banks, fintechs, insurers and retail. Some capabilities, such as scam defense with AI agents, are currently offered in the United States.

  • GET STARTED

Find where your customers are losing money.

Request a Fraud Journey Assessment. We map your customer journey, what each tool covers today and the layers that would close the biggest gaps.

  • Where losses concentrate, stage by stage
  • What your current tools cover and miss
  • A layer architecture and a first PoC
  • No obligation
  • Response within 1 business day
  • NDA-first

Cybersecurity for critical organizations. Mercurius combines Offensive Security, SOC AI, Threat Intelligence, Vulnerability Management, Anti-Fraud and Cyber Governance to help organizations anticipate, detect and neutralize cyber threats before they impact operations.

Solutions

Discover

Defend

Exposure Management

Govern

Fraud Prevention

Hawk

Cactus

Mercurius

Company

Blog

AWS Marketplace

Contact

  • Santiago - Chile
  • São Paulo - Brazil
  • Miami - USA
Linkedin Instagram Whatsapp Envelope

© Mercurius. All rights reserved.