Fraud and AML tools for banks and credit unions: which tool for which job

Which tool for which job?

Your job	Documented example	The reference to read	What it documents
Write detection rules per program manager and partner	Unit21	Patriot Bank	Patriot Bank builds transaction-monitoring rules for each program manager and co-branded partner in Unit21, and handles cases and SAR filing in the same platform. 1
Monitor Pix, card and other transactions in real time	Feedzai	BTG Pactual	BTG Pactual monitors Pix, card and other transactions in Feedzai, with custom rules, device recognition and facial biometrics. 3
Spot fraud from device and behavior signals	Feedzai	Ibercaja	Ibercaja uses Feedzai Digital Trust (behavioral biometrics, malware detection) to see how customers use their devices and adjusts its fraud rules. 5
Verify identity and monitor transactions	Feedzai	novobanco	novobanco uses Feedzai Digital Trust to verify customer identity and Feedzai Transaction Fraud to monitor transactions in real time. 6
Detect fraud across digital sessions	Feedzai	TBC Bank	TBC Bank applies Feedzai Digital Trust to behavioral, device and transactional data across digital sessions to block suspicious activity. 7
Resolve entities for corporate KYC investigations	Quantexa	ABN AMRO	ABN AMRO runs entity resolution and network analytics in Quantexa on client and company-registry data for its analysts' complex KYC investigations. 8

What does the public record cover for banks and credit unions?

As of 2026-11-03, usedby lists 52 Banking & Mortgages companies that use a Fraud & AML tool: 19 are CONFIRMED by a case study, quote or press release, and 33 are OBSERVED on a vendor's own customer page. Only a CONFIRMED source that describes a workflow says what the institution does with the tool. An OBSERVED line shows adoption and nothing more.

This article follows the work rather than the brand. Unit21, Feedzai, Bretton AI and Quantexa each have at least one bank source that names a task. Sardine and ComplyAdvantage appear for banks mainly as customer-page listings, for example Coastal Community Bank and Emprise Bank, and Hawk is covered here only through its catalogue entry. Nothing below ranks one tool above another.

91011121314

Which tools monitor transactions and let analysts write fraud rules?

Patriot Bank, which reports about $974 million in assets and ten locations in New York and Connecticut in its Unit21 case study, moved to Unit21 for transaction monitoring, case management and regulatory filing. Its payments division had grown into reloadable corporate and payroll cards through co-branded partners, and the old system could not be tailored per program. The bank now writes rules for each program manager and partner, and reports fewer duplicate alerts and a simpler SAR evaluation step, without published figures.

Kinecta Federal Credit Union replaced Verafin with Unit21 for transaction monitoring, case management and fraud consortium data. Analysts build rules, test them in shadow mode and publish them across check, ACH and wire activity in one place. The source ties the choice to real-time integration with the core and online banking providers. Unit21 reports daily alerts falling from over 100 to fewer than 50, and $70,000 in fraud stopped in one weekend through the consortium. These are vendor-published figures for one credit union with 270,000 members and $6.8 billion in assets.

BTG Pactual, a Brazilian bank serving high-income clients, uses Feedzai for real-time monitoring of Pix, card and other transactions. The setup includes custom rules, device recognition and facial biometrics, and lets customers switch off features such as card-not-present payments. The story reports low fraud rates with high approval rates but gives no numbers.

What transfers is the pattern: rules owned by analysts, with cases in the same place. Whether scoring is real time on your own core is the criterion to test, because only the Kinecta and BTG sources make it a selection point.

1231516

Which tools prepare investigations and due-diligence reviews?

First Internet Bank uses Bretton AI agents to automate research and case preparation for enhanced due diligence, SBA loan origination, fraud reviews and periodic customer due diligence. The agents compile findings into one report so analysts can concentrate on high-risk areas. Bretton publishes an 87% reduction in manual review times for the SBA loan review; that is the vendor's figure for one workflow.

Grasshopper Bank uses Bretton AI's Diligence AI for enhanced and customer due diligence on business customers. The agents do online research, review documents, aggregate registry data and analyze transaction history, which lets the bank onboard more clients at the same standard. The Chief Compliance Officer says analysts spend their time on risk decisions rather than manual tasks, and the vendor reports 70% less time on the enhanced due diligence step.

The Bretton and Unit21 bank sources differ in which job they document, not necessarily in what the products do: Bretton's are due-diligence and review preparation, Unit21's are rules, monitoring and case handling. They cannot be ranked against each other.

24171819

Where do behavioral and device signals fit in fraud prevention?

Ibercaja, a Spanish group with 2.5 million customers and 893 branches, uses Feedzai Digital Trust to analyze how customers interact with their devices. The context lets the bank adjust fraud rules while limiting friction. The story reports an 80% reduction in fraud losses, but attributes it to a combination of solutions that includes Digital Trust, so it is not a measure of Feedzai alone.

novobanco uses Digital Trust to verify customer identity and Feedzai Transaction Fraud to monitor transactions in real time. TBC Bank uses Digital Trust on behavioral, device and transactional data across sessions. TBC's story gives a 65% share but words it both as fraud cases and as fraudulent sessions, so that figure should not be reused without checking the metric. Raiffeisen Bank International appears as a Feedzai customer at OBSERVED level only.

56720

What do the Quantexa sources actually show?

ABN AMRO started a proof of concept in late 2019 for KYC onboarding of corporate clients. Quantexa ran entity resolution on the bank's client database and company-registry data, and added network analytics. The interface went first to analysts handling complex KYC investigations, and the bank's product owner says they spend less time gathering and interpreting data.

HSBC's Quantexa source is an interview with its Global Data CIO about a single customer view built with entity resolution, which merged four tools into one and carries a potential saving of about £4 million. HSBC calls the rollout early-stage and mentions financial crime only in passing, so it documents a data foundation rather than an AML workflow. Standard Chartered and BNY are OBSERVED with no task. Quantexa lists a cloud AML offer for mid-size and community banks, but no such case is among these sources.

8212223

What do the US peers show, and at which proof level?

Four US institutions have CONFIRMED lines: Green Dot Bank and Kinecta Federal Credit Union with Unit21, and First Internet Bank and Grasshopper Bank with Bretton AI. Coastal Community Bank (Sardine and Bretton AI), First Federal Bank of Kansas City (Sardine) and Emprise Bank (ComplyAdvantage) are OBSERVED, meaning a vendor shows them as customers. HSBC (Quantexa, UK) is OBSERVED too.

For a US bank or credit union, Kinecta, First Internet Bank and Grasshopper Bank are the ones with a described workflow to raise with a peer. Green Dot Bank is confirmed as a Unit21 customer through a named quote, but no task from it is used here. The OBSERVED banks name no task, so ask what they use the tool for before assuming anything. Coastal Community Bank shows two vendors at one bank, and a Bretton quote from its president mentions shorter L1 investigation times, but neither source says how work is split between the tools.

241314181922

What should you check before reusing a documented workflow?

Start with the data path. Kinecta's story names Symitar and Alkami as sources of fragmented data, and Patriot Bank says its old vendor made data extraction hard, yet Unit21 was ingesting data within days. Scale also differs: Patriot reports $974 million in assets and Kinecta $6.8 billion, so volumes behind any figure are not comparable. Both Unit21 and Bretton AI list SOC 2 Type II on their own security pages; that is a vendor statement, not a clearance for your regulator.

Our editorial suggestion, not a vendor capability: replay a sample of your own past alerts and cases through any shortlisted tool and compare analyst handling before committing. Pricing and pilot terms are not shown in these sources, so treat them as unknown. Real-time latency, alert volumes and effects on loss rates are also unpublished for every bank here.

121516172425