Zoot Enterprises

Hosted decision management dates from 1992, and a stateless application processing service and a graphical rules editor for business users still sit at the center. When the selling point is that a business user writes the decision logic, that logic is mostly not learned.

Control sits with the institution and is evidenced from the customer side, with one lender describing changes to the entire origination flow, not just the decision step. A vendor that hands over the logic makes no autonomous decisions on anyone's behalf.

Machine learning and agentic components now sit in a decision path that was deterministic, and nothing published says how they were validated or how they perform. Proven models for risk and profitability are offered with no accuracy figure, benchmark or drift disclosure behind them.

Three European lenders are quoted with specifics, one citing better approval rates without looser controls, but no current United States customer is named. The claim of three of the top five United States banks comes from a 2012 analyst assessment never restated, and no outcome figure is published.

Auditable records and configurable testing tools let a customer evidence its own decisions, which concerns the system rather than the machine learning and agentic parts recently added. Those components have no model card or evaluation method, and whether one institution's application flow informs models offered to others is unaddressed.

Assurance material cites health information privacy regulation and a health sector security certification, odd references for a credit decisioning business. The statute that does govern this work, the Gramm Leach Bliley Act (GLBA), goes uncited, and neither regional site publishes a privacy program, retention schedule or subprocessor list.

PCI DSS certification is held, and SOC 1 and SOC 2 audits recur annually, which answers the currency question a single dated report leaves open. The European site misstates the same credentials, calling an attestation a certification, and no trust center offers reports without contact.

Compliance is sold as a product line, yet the Equal Credit Opportunity Act (ECOA), which governs automated credit decisions in the home market, is never cited. With a separate European entity and European customers, the company is silent on the EU AI Act, under which credit scoring is a designated high risk use.

Business users can change decision rules in a graphical editor without engineering, so a rule with disparate impact can reach live applicants with no fair lending review or technical change control. The company publishes no fairness testing, explainability documentation or expected approval step for rule changes.

The 99.9 percent availability figure describes the facility's capability rather than a service level with credits, remedies or a measurement basis, and no terms, warranty or liability cap is published. An outage halts lending outright, while the rules themselves are customer authored, which puts the substance of each decision with the institution.

Models from third parties can run in the decision path through the provider network, an admission that the platform is partly a conduit. No bureau, identity service or model supplier is named, and the vendor's own machine learning has no stated technique or provenance.

Hundreds of credit, fraud and identity sources arrive through a single gateway, so a lender integrates once rather than per provider. Integration with any core banking or credit union system is promised without a hedge, which matters to smaller institutions, though no developer portal or connector catalog lets anyone verify it.

Processing runs in an owned data center at the Montana headquarters, a seismic rated building on a fiber loop, so no hyperscaler sits between vendor and data. Separate United States and European entities imply European processing, but where European client data actually runs and under what residency commitment is unstated.

A major software directory notes that no pricing has been provided, and none appears elsewhere. Implementation services, workflow embedding and partner data connections each carry costs of a different shape, and nothing shows whether decisions, applications or an enterprise fee drive the bill.

Banks, automobile manufacturers and retailers sit in the stated client base, reaching captive finance and retail credit, with credit unions and fintechs addressed separately. Coverage runs from prescreen and account opening through fraud to collections and recovery across four continents.