GDS Link

Lenders configure their own criteria, rules and scorecards, so the platform makes credit decisions with no model involved at all. Machine learning lives in an optional analytics module, applied by the platform rather than constituting it.

A named credit union automates 65 percent of its decisions, which also discloses that 35 percent still reach a person, a more useful figure than any oversight claim. The lender sets that dial, and the vendor states no view on what should not be automated or what review precedes an automated decline.

Scorecard modeling and a governance module give lenders somewhere to document models, but the vendor publishes no accuracy, benchmark or validation method for its own analytics. Its outcome figures measure throughput, so nothing public shows the models separate good borrowers from bad any better than what a lender already runs.

One credit union, quoted by name, reports application processing falling from three days behind to 45 minutes and revenue more than tripling over five years. That is the only named customer for a company operating since 2006 at hundreds of thousands of decisions a day, with no customer count or analyst placement.

Model governance and post booking behavior monitoring are sold to lenders as features, which says nothing about how the vendor evaluates its own analytics. Whether one lender's application data informs models or benchmarks offered to another is unaddressed, a live question for any multi tenant decisioning engine.

Bureau, verification and alternative data for every application flows through vendor infrastructure, concentrating consumer financial information in one place. The company publishes no privacy program, retention schedule or subprocessor list, and never cites the Gramm Leach Bliley Act despite selling to banks and credit unions.

The company publishes no trust center, named certification or attestation report, and no security page at all. Smaller credit unions and specialty lenders, who have little leverage to demand documents in procurement, carry the cost of that gap.

Credit decisioning sold in the United States sits squarely under the Equal Credit Opportunity Act (ECOA) and its adverse action notices, yet the company cites no governing statute by name. Despite an office in Spain and deployments across several countries, no position on the EU AI Act appears, though it treats credit scoring as high risk.

A partnership with a specialist firm brings in fairness intelligence for lending, a more candid move than claiming the problem solved internally. What that capability does, whether it is generally available and any disparate impact result are all unpublished, though the platform executes declines.

A failure here is a stopped or wrong lending decision, and the company publishes no terms, service level or liability cap covering it. Because lenders configure their own criteria, responsibility for a decision's substance sits more clearly with the institution, though nothing published says so.

Partnerships with risk and data analytics firms are acknowledged as the source of client data, but no bureau, verification service or alternative data provider is named. The analytics module discloses no technique, framework or model provider, nor what data its models were developed against.

More than 200 external sources arrive with attributes already defined, ready for a rule to reference, which saves a lender the mapping work a raw bureau connection leaves. Modules can be adopted separately, but no interface documentation, sandbox or connector catalog confirms whether a buyer's own bureau or origination system is on the list.

Software as a service is the delivery model, with no hosting provider, region or residency commitment published. European customers need processing location documented under data protection law, and nothing addresses it.

Pricing stops at a claim that the solutions fit a wide range of applications and budgets, which places no buyer within it. With a data layer, a decision engine and analytics sold as modules, the charge could run per decision, per module or per seat, and nothing published narrows it.

Banks, credit unions and specialty lenders each get their own published material, with arguments that differ sensibly between segments. Coverage runs across originations, account management and collections from seven international offices, so one decisioning layer can serve functions that usually sit in separate systems.