FraudShield
FraudShield is an anti money laundering transaction monitoring platform from Nexiant, the Australian risk company behind MemberCheck, NameScan and GPayments. It is sold on its own, with its own dashboard and credit packs, to money laundering reporting officers, heads of risk and compliance officers at banks, fintechs, payment providers, remittance businesses and online betting operators.
Transactions arrive through an API in real time or in batch files and run against a rule plan. A default plan of 46 rules is live from day one, and staff can edit rules or describe new ones in plain language with the AI Rule Builder. A rule hit opens a case, which moves through assignment, escalation to the reporting officer and a second reviewer's approval before it closes.
The AI Engine adds a 0 to 100 score to every transaction. Each institution trains its own model on a finalized rule plan, using synthetic transactions generated from those rules rather than customer history, and the score shows which rules pushed it up or down. Near miss detection, off by default, can open a case for a high score that tripped no rule. Report templates prefill case and transaction data for a filing with a financial intelligence unit such as AUSTRAC, FinCEN or the UK's National Crime Agency. Packs start at $2,400 for 10,000 production credits, one credit per transaction.
Capability Axes
Capability grades
15 of 15 axes rated · 10 graded A or B
The rules decide what gets flagged. The AI Engine adds a separate 0 to 100 score that measures how close a transaction came to breaching the plan's rules, and it never blocks processing or changes a rule result. An institution can run FraudShield with no model enabled, and then transactions simply go unscored while the rules engine, case management and reporting carry on as before. The AI earns its place in near miss cases, where a configured score range opens a case no rule would have raised, and in the AI Rule Builder, which turns a plain language description into a rule.
The score never blocks a payment and never overrides a rule. Rules flag transactions and open cases, and an institution can add score ranges that open a Low, Medium, High or Urgent case when no rule fired, a setting that is off by default. Cases are assigned round robin or to the analyst with the fewest open cases, escalate to the money laundering reporting officer and need a second reviewer's approval to close, with an audit log of each step. Any action on the payment itself sits in the institution's own systems.
Every scored transaction shows up to three top rules, each with its impact in points, its share of influence and whether it raised or lowered risk, calculated with SHAP. A finalized plan cannot change its rules, so each model stays matched to the plan it learned. One model runs per plan per environment, a model can be proven in Test before it reaches Production, and a batch that is run again uses whichever model is active at the time. Accuracy, precision and recall figures, and validation material for an SR 11-7 review, are not offered.
Uptime of 99.98% across Nexiant's services is the one figure, with no measurement period, and Nexiant says FraudShield has live clients in production without naming any. Its named customer stories belong to MemberCheck, NameScan and GPayments. Case records keep the rules that fired, the AI score with its rule impacts and the analyst's disposition, so an institution can measure the model's effect after rollout. FraudShield reports no detection lift, false positive rate or investigation time saved for any customer.
Models train on synthetic transactions generated from the institution's own finalized rule plan, so customer transaction history never becomes training data and nothing is pooled across clients. Each model stays bound to its frozen plan and scores only while that plan is active in the environment where it was enabled. Rule values from live transactions feed the model as inputs, not as training examples. Retention of trained models, and their deletion when an account closes, are not addressed.
Each transaction can create or update a customer profile linked to that customer's accounts, past transactions and cases, so the product holds personal and payment data. Test and Production keep their transactions, profiles, cases and batch uploads apart, while rules and users are shared. Nexiant lists GDPR and DORA beside its ISO 27701:2019 privacy badge, but its public privacy policy covers the website, and the product terms customers accept sit inside the account. A retention period for transaction records and a list of the processors handling them are not available before sign up.
The product page lists ISO/IEC 27001:2022 and ISO/IEC 27701:2019 and calls FraudShield's infrastructure independently certified and audited. The certifier, certificate dates and scope are not given and no certificate opens from the site, while Nexiant's about page describes the group's controls as aligned with ISO 27001:2022. Inside the product, roles control who can manage billing, train models or view rules, each API key belongs to either Test or Production, and case management keeps an audit log.
FraudShield supports the institution's own reporting duty. Compliance staff flag a case for reporting, prefill a configurable template with case and transaction data and export the report as XML, CSV or PDF, and saved reports cannot be edited or deleted. AUSTRAC, FinCEN and the UK's National Crime Agency are the intelligence units it is built around, and Nexiant aligns the product with FATF guidance, AUSTRAC rules and EU frameworks. Submission stays with the institution, since FraudShield does not file directly with any authority.
Permissions separate who can view rules, train models and change the live plan, and only finalized plans can be trained. A trained model must be switched on for a specific environment before any transaction gets a score, and each environment runs its own active plan. Score ranges for automatic cases are set by the institution, and every case records how it was resolved. FraudShield offers no fairness evaluation of how scores fall across customer groups, a gap that matters because profiles tie a person's accounts and history together.
An investigator can trace any score to the transaction behind it and the rules that moved it most, and that detail travels into case records and exported reports. Staff can open a case by hand when neither rules nor score ranges did. The product terms sit in the customer's account, and FraudShield offers no accuracy commitment or compensation for a wrong score. Case review is an internal compliance workflow, not an appeal route for the customer whose transaction was flagged.
Each institution's model is built inside FraudShield. A pipeline generates synthetic transactions from the finalized plan, extracts each rule's values and trains a model on them within Nexiant's own AI Engine, which keeps the training chain to the institution's rules rather than a shared data set. The model family, any outside framework or model provider, the language model behind the AI Rule Builder and the cloud that runs training and scoring are not named.
Payment systems send transactions one at a time through the API or in batch files, with an environment key routing each to Test or Production. A transaction carries identity, account and payment fields, updates the customer's profile and is checked against the plan's rules on arrival. The API returns an AI status with each response, NotEnabled or Error when no score is available. A ready built connector to a core banking platform, payment processor or outside case management system is not part of the product.
FraudShield runs as a hosted service with separate Test and Production environments, each with its own transactions, customer profiles, investigations, batch uploads and API credentials, while rules, users and audit logs are shared, so an institution can trial a new plan or model beside live monitoring. The hosting location and cloud provider are not stated, and buyers have no choice of region or an on premises option.
Credit packs are priced in public. $2,400 buys 10,000 production credits and 1,000 test credits, about $0.24 a transaction, and $4,800 buys 25,000 production and 2,500 test credits, about $0.19 each. The first package carries a $500 onboarding fee, credits expire after a year, and custom packages are quoted on request, with other currencies at checkout. One credit covers one transaction, by API or batch. A free trial needs no credit card and comes with test credits and the 46 rule default plan, and billing managers can top up by hand or set automatic replenishment.
FraudShield is sold to money laundering reporting officers, heads of risk and compliance officers who need automated monitoring under anti money laundering and counter terrorism financing rules. Banks, fintechs and other regulated financial institutions are the core buyers, with dedicated workflows for payment service providers with shifting volumes, remittance businesses and online betting operators.
Nexiant's industry pages extend the pitch to insurance, cryptocurrency, legal and real estate firms. The same rules engine, case management and reporting serve every segment, with rules tuned to the fields that matter in each.
Alternatives to FraudShield
Other products that do the same kind of work as FraudShield, and what each one does.
Tookitaki unifies anti money laundering monitoring, fraud prevention, screening, onboarding risk, customer risk scoring and case management in FinCense, a platform for banks, digital banks, digital wallets, remittance firms and payment institutions across Asia Pacific.
Eastnets sells compliance and payments software to banks, payment firms, exchange houses and insurers.
Vneuron sells the Reis Risk and Compliance Suite, an AML and counter terrorist financing platform for banks, insurers, leasing companies, investment funds, brokers, money transfer businesses and microfinance institutions.
Consilient sells anti money laundering and counter terrorist financing models built with federated learning, so banks, regulators and financial intelligence units can learn from each other's behavioral patterns without pooling their data.
Flagright sells a financial crime compliance platform to payment processors, digital banks, banks, credit unions, brokerages, crypto and remittance firms, covering real time transaction monitoring, dynamic customer risk scoring, sanctions and watchlist screening, case management and…
T-NEX Fraud Detection is a transaction monitoring and case investigation module for fraud and anti money laundering teams at regulated institutions.
Pricing
Vendor-published figures are labeled as such. Figures labeled “Estimated” are derived from third-party sources and have not been confirmed by the vendor.
| Entry Price | Pricing Basis | Data Protection Terms | Implementation | Source |
|---|---|---|---|---|
|
$2,400 for 10,000 production credits
$2,400 baseline
|
One credit covers one transaction, sent by API or batch file. $2,400 buys 10,000 production and 1,000 test credits, about $0.24 a transaction, and $4,800 buys 25,000 production and 2,500 test credits, about $0.19 each. Credits expire after one year. | Test and Production keep their transaction data apart, and each organization manages its own users and roles. Nexiant lists GDPR and DORA alignment, and the product terms customers accept sit in the account, apart from the website privacy policy. | The first package carries a $500 onboarding fee. A new organization can start a free trial without a credit card, invite staff and configure Test before buying Production credits. | Vendor Published |
FraudShield is sold in prepaid credit packs, one credit per monitored transaction, starting at $2,400 for 10,000 production credits. Custom packages are quoted on request, prices are in US dollars with other currencies at checkout, and billing managers can top up by hand, set automatic replenishment or buy by invoice.