EverC

The removal test leaves manual merchant review, which cannot operate at the volumes involved. Models classify billions of product level data points spanning text, images and metadata, scanning more than 30 million items each day, while a proprietary risk graph reads relationships between web addresses to identify merchants reappearing under new names. Category code classification is automated at onboarding. Determining that a listing is counterfeit or that an apparently unrelated merchant is a returning bad actor is inference throughout.

The tooling is built around compliance teams making decisions rather than around automated action, with dashboards ranking marketplace risk so teams can prioritise attention, findings customisable for the customer's own thresholds, and a workflow for tracking mitigation requests through to resolution. That last feature is telling, since it assumes a negotiation with the marketplace or seller rather than unilateral removal.

Against it sits the company's own framing that the solution detects, identifies and removes high risk merchants, and nothing reconciles automated removal with the review workflow or states which decisions require a person.

One published figure is both specific and the correct metric: over 95 percent precision across more than 78 million listings at a named customer. Precision rather than recall is the right measure here because the cost of the error falls on legitimate sellers wrongly removed, so a vendor choosing to publish it is answering the question that matters to the assessed party rather than the one that flatters the product.

Volume at that scale also makes the figure meaningful rather than anecdotal. What is absent is recall, since nothing states what proportion of illicit listings the system misses, and no accuracy figure exists for merchant classification as distinct from product scanning.

Two customers are named with results attached: a major online marketplace reporting more than 78 million listings scanned in the first months of the partnership at over 95 percent precision, and a global payments platform describing the work as instrumental to its ongoing merchant monitoring. Throughput is stated at over 30 million items daily. A major financial information provider profiled the company and its products in a market intelligence report.

The 2025 merger with a merchant risk and compliance firm is described as producing a combined business serving most major payment providers globally, spanning banks, acquirers, marketplaces and online platforms. The company has operated since 2015.

Cross customer intelligence is disclosed as the mechanism rather than hidden, since the risk graph explicitly surfaces repeat offenders and known bad entities, which only works because observations accumulate across the whole customer base. That is the honest form of pooling: a payment provider adopting the product understands it is both contributing to and drawing from a shared view of who has been removed before. Held at B because nothing states what each customer contributes, whether a merchant's flagged status at one acquirer is visible to another, or how long entity associations persist in the graph.

No data protection agreement, retention schedule, subprocessor list or deletion commitment was located. Exposure is structurally lower than for consumer facing vendors because the subjects are businesses and much of the input is publicly posted product and website content, though merchant onboarding necessarily involves business verification touching identifiable owners, and the risk graph retains associations between entities over time. Nothing describes how long a merchant remains linked to a prior association.

No attestation, certification, trust centre or enumerated framework was located. Banks and major payment providers are the named buyers and their supplier assessment programmes gate anything touching merchant onboarding, so the underlying assurance has been examined repeatedly in private while nothing is published for a prospective customer to read.

No regulator, statute or rule is named. Two regulated concepts appear correctly, the risk based approach that anti money laundering supervision requires of obliged entities, and merchant category codes as the card network classification whose misuse constitutes transaction laundering. The company also states its products help marketplaces maintain compliance with regulatory changes without identifying any. For a business whose entire purpose is helping regulated payment providers meet obligations they hold, naming those obligations is the obvious missing disclosure.

The stakes for the assessed party are severe and undiscussed. A merchant classified as high risk loses payment acceptance, which for an online seller is not a inconvenience but the end of the business, and the decision is made by a third party the merchant has no relationship with.

The risk graph compounds this by flagging suspicious associations between web addresses, so a merchant can be caught by connection rather than conduct, which is the same proximity problem recorded elsewhere in this index for transaction graph analysis. Category based monitoring adds a further layer, since operating in a category the model treats as high risk attracts scrutiny regardless of individual behaviour. No false positive rate, appeal process or reinstatement path is published.

No guarantee, indemnity or correction process was located for the vendor's own errors. The customer is served by a mitigation request workflow that tracks issues through to resolution, which is a real mechanism. The merchant is not: a seller removed from payment acceptance or delisted from a marketplace deals with the platform rather than with EverC, may never learn a third party classification drove it, and has no described route to contest a category assignment, a graph association or a product level determination, or to have a correction propagate to other platforms drawing on the same intelligence.

The risk graph and underlying datasets are described as proprietary, which establishes ownership, and no model provider, hosting arrangement or external data supplier is named. The material dependency for a product scanning 30 million items daily is access to marketplace listing content, and nothing describes how that access is obtained, whether through partnership, interface or collection, which determines both coverage and durability. No subprocessor list appears.

Deployment friction is deliberately minimised and stated in concrete terms, with marketplace assessment operating without any integration at all, onboarding insight returned into existing workflows in under fifteen seconds, and category code prefill reducing manual work inside the customer's own application process. For payment providers where onboarding speed is a competitive variable, integrating at that latency is the requirement. No named payment platform, acquiring system, marketplace or onboarding tool appears, and no developer documentation was located.

No hosting provider, region selection, residency commitment or private deployment option was located. The customer base is explicitly global and includes banks and acquirers operating under different supervisory regimes, several of which impose expectations about where merchant and onboarding data is processed, and none of that is addressed.

No pricing, packaging or basis of charge was located across five distinct products. Adoption cost is addressed repeatedly and credibly, with the marketplace assessment product described as requiring zero technical lift and no integration, and onboarding insight returned in under fifteen seconds, which for a payment provider weighing merchant acquisition speed against risk is the operative constraint. What any of it costs is not stated.

Buyer coverage spans banks, merchant acquirers, payment providers, marketplaces and online platforms, which is the whole chain that carries liability for what merchants do, and the post merger claim is service to most major payment providers globally. Functional coverage runs from onboarding through lifecycle monitoring to product level scanning and whole marketplace assessment.

The limit is subject: this is merchant and marketplace risk specifically rather than the wider financial crime stack, and the deepest expertise sits in ecommerce rather than in banking more generally.