Fourthline

Document authenticity, liveness and fraud detection run on models, and they drive verification. Remediation of existing records, screening and qualified electronic signatures round out a compliance platform with value beyond the models.

Human experts review cases through a case review portal, and in Germany and Austria regulated video identification puts a live agent in the session by law. When a case escalates from automated to expert review, and how the two resolve disagreement, is not described.

Approval accuracy is published as a figure, 98.7 percent, and writing on document level variation explains effects such as where security features sit on particular national identity cards. The method behind the figure, error analysis by market and a documentation package for a bank's own validation are absent.

Revenue of about $56.7 million and profitability are disclosed, and an analyst study reports 15 to 20 minutes saved per case and a 25 percent conversion gain at one fintech bank. Neither the study's author nor the bank is named, and the homepage joins no customer to a result.

Insights on service performance train IT systems that may include AI, and identity pictures are converted into vectors, so verification data feeds model training across the customer base. Whether one institution's verifications shape what other clients receive, or whether a customer can opt out, is not stated.

Onboarding data, including biometrics, is set out with its legal ground in a privacy statement available in more than 18 languages. Retention runs five years after the relationship ends, and up to eight years for fraud related data.

EY CertifyPoint issued the ISO 27001:2022 certificate, number 2021 039, which a buyer can check with the issuer, and an ISAE 3000 assurance report and iBeta Level 1 and Level 2 presentation attack testing are listed. A trust portal, downloadable report or SOC 2 report is not linked.

BaFin and the Deutsche Bundesbank reviewed the due diligence flow and screening at the request of business partners under examination, a regulator examining the product rather than a vendor asserting compliance. Fourthline Payments, a group entity, acts as controller for client onboarding, though no formal admission with published criteria exists.

Racial bias appears in an August 2024 newsroom piece only as an industry challenge in sourcing data. Demographic pass rates for Fourthline's own face matching and liveness, an independent demographic evaluation and accessibility analysis are all unpublished.

Individuals can have incorrect personal data rectified, and remediation tooling lets an institution correct records it later finds wrong. An applicant who fails verification has no described way to learn why or ask for reconsideration, and no accuracy guarantee or remediation term stands behind the outcome.

Amazon Web Services, Google Cloud and Genpact appear in a public table of processors with purpose and location, and the analytical layer is proprietary, described as a sovereign technology stack. Watchlist and screening data suppliers are not named.

One API integration opens the full modular set, from verification and screening through remediation and electronic signatures, with documentation at docs.fourthline.com and AWS Technology Partner status. Which core banking or onboarding platforms it already connects to is not said.

Processors run in Ireland, Germany and Romania among other locations, with adequacy decisions, the EU US Data Privacy Framework or standard contractual clauses covering transfers. Customer selectable regions and an on premises option are missing from the privacy statement.

A sales contact is the only route to pricing, with no rate, tier, billing unit or minimum published. Its modular structure makes scope the first thing to settle.

Banks, fintechs and insurers sit among named institution types across more than 30 markets, with depth in each. Southern Europe, the United States and Latin America arrive through the announced Veridas merger rather than existing operations.