Feedzai

Service Overview

Feedzai is a Portuguese financial technology company positioning itself as an AI-native fraud and financial crime prevention platform. Its core is not a single model but a platform called RiskOps, which places fraud and financial crime governance in one framework across the customer lifecycle - onboarding, transactions, account tenure and suspicious activity reporting.

The platform splits into 4 capabilities: Farol is a resident agentic AI agent that can launch investigations and take action on its own; Feedzai IQ provides actionable network intelligence, turning relationships between entities into usable decision input; Orchestration handles the account-opening flow; Advanced AI is the underlying model capability. Stacked together, they show the product shape is not "hand back a score" but "take over a segment of the process".

The solution list is long, 13 areas in 3 groups: on the fraud side, omnichannel fraud, transaction fraud, new account fraud, scam prevention and account monitoring; on the compliance side, AML, AML transaction monitoring, watchlist screening, identity verification and secure onboarding; on the institutional side, risk management for acquirers and Merchant IQ Signals. Keeping those three groups on one platform is the most important property of the product - in real cases, fraud and money laundering are frequently two ends of the same funds chain, and running them in separate systems breaks the evidence chain and duplicates reporting.

The site ships in 4 languages - English, Spanish, Brazilian Portuguese and French - for multi-region delivery at multinational institutions.

Its audience must be stated plainly: banks, acquirers, payment institutions and licensed financial entities. It is not tooling for an e-commerce checkout page. For merchants the capability is surplus - running a bank-grade engine against a small scenario means configuration, tuning and operating cost exceed the benefit. This matters because capability lists are seductive, and the question people forget is who carries the configuration cost.

Key Strengths

• Fraud and AML governed on one platform: in real cases skimmed funds are laundered through layers of accounts, so if fraud decisions and AML monitoring live in separate systems the same funds chain splits into two evidence sets. One platform keeps relationship analysis on a single chain and aligns regulatory reporting terminology.
• Agentic capability takes over the investigation process: Farol's value is not the decision but the action after it - pulling evidence, linking entities, executing dispositions. The real cost centre in fraud control is investigation and execution, not scoring, so this automation hits headcount directly.
• Onboarding is covered: new account fraud and secure onboarding are gaps in many risk systems. When an account is created on a false identity, every later transaction judgement rests on a false premise, which is a structural defect.
• A dedicated acquirer offering: an acquirer manages merchant onboarding, merchant transaction quality and must prove monitoring to card networks. Merchant-side signals like Merchant IQ Signals fit exactly that position instead of treating merchants as anonymous transaction sources.
• Multi-language, multi-region deployment: 4 language versions are not a translation exercise but evidence of local delivery in Latin America and Europe, letting regional teams at multinational institutions use one platform.
• Completeness from a financial crime angle: AML transaction monitoring, watchlist screening, identity verification and scam prevention sit in one solution set, so institutions need not buy each regulatory requirement separately and end up with fewer seams and lower integration cost.

Product Ecosystem

RiskOps Platform

The unified console and decision layer that holds fraud, identity, AML and account monitoring in one framework, acting as the runtime base for every other module.

Fraud

Covers omnichannel fraud, transaction fraud, new account fraud and account monitoring, deciding in real time at onboarding and at transaction time with case management and investigation workflows.

Identity

Covers secure onboarding and identity verification, answering whether a new customer is who they claim to be, sharing the entity relationship view with fraud judgement.

AML

Includes AML transaction monitoring, watchlist screening and compliance automation, aimed at institutions that must prove to supervisors that monitoring is effective.

Scam Prevention

Targets social-engineering scams where the victim authorises the transfer and the transaction looks completely normal, making rule-based detection nearly useless and requiring behavioural and network-level judgement.

Farol and Feedzai IQ

Farol is the resident agent that launches investigations and executes dispositions; Feedzai IQ turns entity relationship networks into actionable signals. Together they cover what happens after a decision.

Risk Management for Acquirers

Merchant onboarding, transaction quality monitoring and merchant-side signals for acquirers, used to manage portfolio risk rather than individual transactions.

Limitations

• Over-capability for merchants: the product is designed around regulated institutions, while a merchant checkout needs only the moment of approve or decline. Running a bank-grade platform against a small scenario leaves the cost structure completely mismatched and the benefit eaten by operating cost.
• The two error directions are amplified on the institution side: a miss in banking can mean supervisory penalties and reputational damage, a false decline can mean complaints and attrition. It delivers judgement and both consequences sit with the institution, with no financial indemnity.
• High data and compliance cost: transaction, identity, watchlist and relationship data is required, and cross-border deployment falls under GDPR; exporting personal data from mainland China also implicates the Personal Information Protection Law and the applicable export rules. Compliance and security review for such projects is usually measured in quarters.
• Needs a dedicated team to keep running: rules, models, thresholds and case workflows all need maintenance. Institutions lacking dedicated risk and data engineering staff tend to stay at the initial configuration, with results decaying over time.
• Long procurement and go-live cycles: pricing and implementation costs are not published, and the process involves multiple security assessments, compliance reviews and integration testing. For teams wanting to be live in two weeks, this is not the right choice.
• Does not solve chargeback recovery: it covers fraud judgement and case management but absorbs no chargeback loss and offers no recovery-based service, so after-the-fact loss mitigation needs a separate arrangement.

Use Cases

• Banks and card issuers (★★★★★): onboarding, transactions, account monitoring, AML and watchlist screening are hard supervisory requirements, and one platform addresses both capability and evidence chain.
• Acquirers and payment institutions (★★★★★): managing merchant portfolio risk and proving monitoring effectiveness to card networks is exactly what the acquirer offering addresses.
• Fintechs with licensing obligations (★★★★☆): institutions carrying AML and transaction monitoring duties pass review more easily with one platform than with several point solutions stitched together.
• Money-side operations at large cross-border sellers (★★★☆☆): worth considering once risk complexity approaches that of a financial institution, but usually only necessary where a clear AML obligation exists.
• Mainland China operations (★★☆☆☆): theoretically usable when a mainland entity runs cross-border business, but any export of personal data from the mainland must follow the Personal Information Protection Law and the applicable export rules, so a compliance assessment comes first. Purely domestic business does not need overseas platforms of this kind and has domestic compliance technology without an export step.
• Ordinary e-commerce merchants (★☆☆☆☆): least suitable. With no AML obligation, no onboarding flow and no account system, most modules go unused and the return on investment collapses.
• Small teams needing to go live in two weeks (☆☆☆☆☆): least suitable. Implementation and compliance review run in quarters, and the team size cannot support ongoing operations either.

Pricing

Main cost drivers are monitored transaction volume, account scale, how many modules are enabled, deployment model (cloud or on-premise), whether model tuning and dedicated support are included, and the compliance review workload. Feedzai does not publish standard pricing; headings and final rates follow the current official quote and contract.

FAQ

• Is Feedzai suitable for an e-commerce merchant?
  Generally not. Its design premise is a regulated institution with AML obligations, an onboarding flow and account tenure management. A merchant needs only a checkout decision, and this platform's configuration and operating cost far exceed the benefit; look at lightweight payment risk services instead.

• How does it differ from acquirer-focused risk providers?
  The object of judgement and the scope of responsibility differ. Acquirer risk providers usually answer only whether a transaction should pass, while Feedzai also covers onboarding, account tenure, AML monitoring and watchlist screening, delivering process and evidence chain rather than a single decision.

• What does an agentic capability like Farol actually solve?
  Execution after the decision. The bulk of fraud control cost is investigation - pulling evidence, linking entities, writing conclusions, executing dispositions - work that used to be staffed manually. Agents automate part of it and directly compress headcount. It does need clean entity data and clear disposition authority, or automated execution amplifies errors.

• We only acquire cards and carry no AML obligation. Do we need it?
  No. The test is simple: do you need to prove to a supervisor that your AML and transaction monitoring are systemic? If yes, it is one of the standard options; if not, it sits idle while you carry the data integration and compliance review burden.

• Can a mainland China entity deploy it?
  Cross-border business can be evaluated, but data compliance is a precondition. Exporting personal information must follow the Personal Information Protection Law and the applicable export rules, so confirm whether a security assessment, standard contract or certification path applies. Purely domestic business is better served by domestic compliance technology with no export step.

Source: original 16IDC content. Product capabilities, supported countries and fee definitions follow the official documentation and your contract.