Every transaction scored
Every near miss caught
FraudShield AML transaction monitoring scores every transaction from 0 to 100 and opens a case when a near miss slips past your rules.
€4,200
Direct debit
$9,850
Domestic wire
£12,500
SWIFT
FraudShield
Rules
No rule flagged
AI risk score
82
of 100
Cleared
Case opened
Near miss
Flagged by a rule
No credit card required. Free test credits included.
Platform
From first rule to regulator report
Build rules, investigate cases and report to regulators in one place, with 46 default rules ready on day one.
AI Rule Builder
Describe a rule in plain language and FraudShield writes it for you. Review it, adjust it and add it to a plan.
AI Rule Builder
Flagfiveormoretransfersover9,000in24hours
- Number of transfers
- 5 or more
- Amount
- Over 9,000
- Within
- 24 hours
- Risk level
- High
Test before you go live
Run a plan in the Test environment first. Finalised plans are locked, so what you tested is what runs.
Investigate in one place
Auto-assign new cases, escalate them, and require a second reviewer to close. Attach evidence and export a PDF.
- Open
- Investigating
- Escalated
- Closed
Regulatory reports
Build report templates, flag cases for reporting and export in the format your regulator expects.
- XML
- CSV
API and batch
Send one transaction at a time by API or upload batches by file or API. Map your own field names once.
Control and audit
Role-based access and multi-factor sign-in. An audit log records every change, including AI model training and activation.
- Roles
- Multi-factor
- Audit log
- Rule edited
- AI model trained
- Model enabled for Production
AI Engine · New in FraudShield
Rules decide, AI shows what they miss
A transaction can sit just under every threshold and pass every rule. The AI Engine scores each one from 0 to 100, so you see how close it came and can act on it.
Every transaction, two signals
Each dot is a transaction, placed by its AI risk score. Rules and the AI score work independently.
Flagged by a rule
The rule opens the case, with the AI score stored alongside it.
No rule, score in an enabled range
FraudShield opens a case at that range's risk level.
No rule, score outside the ranges
No case is opened.
Example settings: High and Urgent switched on. Automatic cases are off until you enable them. You choose the risk levels and the score range for each. Defaults are Low 0 to 25, Medium 25 to 50, High 50 to 75 and Urgent 75 to 100.
Train
Pick a finalised rule plan and FraudShield trains a model on it.
- Generate synthetic data
- Extract features
- Train model
Score
Every processed transaction gets a score from 0 to 100.
Act
A transaction that triggers no rule but scores in an enabled range opens a case.
- Risk level
- Urgent
- Rules triggered
- None
- Created by
- System
Rules still decide. If the AI is unavailable, transactions are processed on your rules alone.
Every score comes with its reasons
See which rules pushed a score up or down, ranked by impact, on every transaction and case. Explanations use SHAP, an industry-standard method.
AI Risk Assessment
ExampleTransaction
TX-48201
Rules triggered: none
AI risk score
82/100
Top rules by impact
- +18
Velocity, 24 hours
Increased risk
- +11
Jurisdiction risk
Increased risk
- -9
Account age
Decreased risk
Why it matters
Built for the team that reviews the alerts
Less time sorting, more time deciding.
Review the riskiest cases first
Sort and filter cases by AI risk score, so the transactions that came closest to your rules rise to the top.
- Case 104134
- Case 104291
- Case 104358
- Case 104476
Example
Explain every decision
Case reports include the AI risk assessment and the values behind each rule, with the full transaction details.
- AI risk assessment
- Values for each rule
- Transaction details
Share the work evenly
New cases are assigned automatically, round robin or to whoever has the fewest.
Free trial
Try FraudShield before you commit
No credit card and no obligation. Run test transactions through the default rules, then build your own.
Sign up
No credit card required.
Get test credits
Free credits to test your own use case.
Run the default rules
46 default rules are active from the start.
Build and test your own
Write rules, group them into a plan and test it first in the Test environment.
Includes full access to the API documentation and Help Centre support.
Pricing
Pay as you go, with no lock-in
One credit screens one transaction. Choose a package, or talk to us about volume.
Custom packages
Tailored solutions
Volume discounts available
Payment options
Alternative and offline payment options
$500 USD onboarding fee
Applies to the first package purchase only.
1 credit, 1 transaction
Each credit screens one transaction through FraudShield.
1 year expiry
The oldest credits are used first.
Prices in USD
Other currencies are available at online checkout.
Want to try it first? Start a free trial
Why Nexiant
One platform, four specialist products
Financial crime is constantly evolving, and transaction monitoring needs to evolve with it. We're building FraudShield to help compliance teams respond to emerging risks with the adaptability, precision and control needed for what comes next.
Related Solutions
Pair it with customer screening
Screen customers at onboarding and keep screening them, alongside every transaction.
MemberCheck
PEP & Sanctions Screening
Screen customers and entities against global PEP lists, sanctions watchlists, and regulatory databases in real time.
Learn moreMemberCheck
AML/CTF Compliance
Meet AML and counter-terrorism financing obligations across multiple jurisdictions with a single, policy-aligned compliance platform.
Learn moreMemberCheck
Adverse Media Screening
Surface negative news and reputational risk signals from global media sources, automatically, at onboarding and throughout the relationship.
Learn moreWhat is transaction monitoring in AML compliance?
Transaction monitoring is the ongoing review of customer transactions, including payments, transfers, deposits and withdrawals, to spot patterns that may indicate money laundering, terrorist financing, fraud or other financial crime. Regulated institutions are expected to run a monitoring programme under FATF Recommendation 10 and the national AML/CTF laws that implement it.
How does the FraudShield AI Engine work?
You train a model on one of your finalised rule plans, then enable it for the Test or Production environment. From then on, every transaction processed against that plan gets an AI risk score from 0 to 100 that shows how close it came to breaching your rules. Each score comes with the rules that pushed it up or down.
What is the AI model trained on?
On synthetic transactions generated from the rules in your plan. It does not learn from your customers' transaction history, so it scores how close activity comes to your own rules rather than looking for patterns outside them.
Does the AI replace my rules?
No. Rules still decide whether a transaction is flagged. The AI risk score is a separate signal that runs alongside them, and it never blocks a transaction. If the AI Engine is unavailable, transactions are processed on your rules alone.
What happens when a transaction scores high but triggers no rule?
If you switch on automatic case creation, FraudShield opens a case for it at the risk level of the score range you chose (Low, Medium, High or Urgent). The setting is off by default. Transactions that trigger a rule open their cases through the rules as normal.
How does FraudShield help with Suspicious Activity Reports (SARs)?
A Suspicious Activity Report is filed with a financial intelligence unit, such as AUSTRAC in Australia, FinCEN in the US or the NCA in the UK, when an institution suspects proceeds of crime or terrorist financing. FraudShield lets you build regulatory report templates, flag cases for regulatory reporting and export reports as XML, CSV or PDF.
How does FraudShield support the MLRO and the review team?
An MLRO (Money Laundering Reporting Officer) decides whether to file an external report. In FraudShield, cases can be assigned automatically, escalated, and held until a second reviewer approves the closure when you enable that setting. Evidence can be attached to the case and every change is recorded in the audit log.
Can I try FraudShield before I buy?
Yes. The free trial needs no credit card and includes free test credits. A default rule plan of 46 rules is active from the start, so you can send test transactions straight away, then build your own rules and test them in the Test environment.
Start catching near misses today
Try FraudShield free, or talk to us about your monitoring programme.