Seven engines.
One explainable record.

Merchant Discovery begins with the identifiers a payment provider already has—such as a business name, address, phone number, email, MID descriptor or known URL—and searches for the merchant’s broader public digital footprint. It is intended to connect storefronts, alternate domains, social profiles and related commercial properties that may not appear in the original application.

Underwriting and monitoring are only as complete as the URLs being reviewed. If a merchant operates through an undisclosed site, changes brands, or routes customers to another storefront, a review limited to the submitted domain can miss the business that is really generating transactions.

Website Intelligence converts merchant pages into a structured commercial record. It is intended to crawl accessible content, identify products and services, interpret business-model signals, and capture important policies and customer-facing claims. The engine preserves source context so an analyst can see the language that produced an observation.

A corporate registration rarely explains the actual online offer. Payment risk often appears in product catalogs, checkout flows, refund terms, fulfillment promises, subscription language or marketing claims. Understanding the website is necessary to understand the transaction.

Compliance AI evaluates structured website observations against a maintained risk taxonomy. It is intended to identify prohibited or restricted products, problematic claims, missing disclosures and other indicators that warrant review. Findings should state the observed fact, the relevant policy category and the confidence—not present an unexplained verdict.

Rules differ by product, geography, acquiring policy and card network. Manual reviewers cannot consistently compare every page against every applicable rule at portfolio scale. A policy-aware engine helps prioritize attention while leaving final interpretation and enforcement with qualified people.

Identity / KYB is intended to connect the submitted legal entity with the people, brands and operational signals visible online. It compares application data with public website disclosures and other permitted sources, highlights inconsistencies and keeps unresolved identity questions visible for human follow-up.

A valid company record does not prove that the entity controls the website being underwritten. Mismatched operators, hidden beneficial relationships and recycled contact information can indicate impersonation, front companies or attempts to return after termination.

The Transaction Laundering engine is intended to combine relationship-graph signals, merchant history and investigator review to surface undisclosed businesses that may share payment access with an approved merchant. It distinguishes weak automated leads from stronger suspected relationships and, where a service process supports it, analyst-confirmed findings.

A merchant may appear acceptable at onboarding while processing for a prohibited or higher-risk business behind the scenes. This exposes acquirers and payment facilitators to losses, regulatory scrutiny and card-network action. Detection requires looking beyond the approved URL and following the surrounding network.

Continuous Monitoring is intended to revisit merchant properties on a defined schedule, compare current observations with prior evidence and identify changes that matter to risk. Rather than treating every edit as an alert, it classifies changes and prioritizes those affecting products, business model, ownership signals, policies or connected domains.

Approval is a point-in-time decision; the merchant is a moving target. New products, altered refund terms, ownership changes or a newly connected domain can change the risk profile long after underwriting. Persistent monitoring closes the gap between periodic reviews.

Risk Decisioning brings the other six engines into a configurable decision layer. It is intended to combine customer policy, finding severity, confidence and merchant context into a score and an approve, review or decline recommendation. Each factor remains visible so teams can challenge, override and document the result.

More signals do not automatically create better decisions. Teams need consistent prioritization, defensible reasoning and a record of how a recommendation became a final disposition. Explainable decisioning reduces review friction without pretending that a model owns the compliance decision.