Approval Screen is a scaleable, reimagined system for delivering loan servicer approvals of property owner decisions. Approval Screen replaces vague contractual consent standards with objective, auditable algorithmic decisions, delivering near-instant approvals backed by rigorous financial analysis. Our MVP focuses on lease approvals.
Commercial mortgage agreements require lender consent for leasing actions, but the governing standards are imprecise, subjective, and unscalable.
Lack of objective criteria means unpredictable outcomes. Delays hamper property management operations. Processing fees compound frustrations.
Do we agree to vague reasonableness standards for needed decisionmaking approvals, knowing that they limit our practical ability to disapprove of decisions that hurt us? Or do we subject a valuable future customer to an approval delivery system they hate?
Reputational and economic damage spikes when market forces cause request volumes to surge. Manual review can't scale, and decisions remain non-reproducible and poorly auditable.
CMBS servicers have a bad reputation as a group. The tight strips they earn are fixed at origination, creating a perverse incentive to limit decision analysis. Cross-institution knowledge management is absent. Decision logic is opaque. Marketing disadvantages translate to fuller loans, lower coupons, flatter amortization, and more risk in subordinate tranches.
A structured, six-stage pipeline that ingests a lease document and outputs an auditable approval, conditional approval, or disapproval with written rationale.
The borrower selects the request type, new lease, assignment, sublease, termination, or modification, via a hierarchical interface that isolates the applicable decision logic.
Request Type Isolation ModuleThe uploaded lease document is processed by an AI extraction engine that classifies and populates structured data fields across six categories: rents, deposit, expenses, required provisions, prohibited provisions, and lender policies. Governed AI facilitates human decision-making to improve algorithms continuously.
Lease Data Extraction EngineExtracted lease data is cross-referenced against tenant credit information, existing property leases, pre-approved forms, and lender policy databases to surface conflicts and covenant issues before submission.
Relational Data LayerPrior to algorithmic evaluation, the borrower reviews all extracted fields and certifies data accuracy, along with specific representations such as affiliate status and radius clause compliance.
Borrower Certification InterfaceThe system applies sequential approval gates, DSCR analysis, LTV severity evaluation, prohibited and required provision checks, lender policy compliance, and tenant improvement caps, generating an approval determination with full rationale.
Decision Algorithm EngineAn automated narrative engine generates a written servicing decision citing the applicable servicing standard and delegation agreement, producing an audit-ready explanation document with every determination.
Automated Narrative EngineEvery lease passes through a defined sequence of approval tests. A single failure at any gate produces a disapproval. All gates satisfied yields automatic approval.
Compares pre- and post-transaction Debt Service Coverage Ratios across all loan years. Any post-transaction DSCR below 1.00 triggers disapproval.
Where DSCR decreases, an incentive-aligned deference factor (1 to 5) is assigned based on LTV thresholds, establishing a minimum DSCR floor between 1.05 and 1.25.
Evaluates pre- and post-transaction Loan-to-Value ratios at inflection points. LTV increases beyond a permitted threshold, derived from an exponential function curve, trigger disapproval.
Identifies concentrated default risk periods, scheduled vacancies, capital expenditure demands, and loan maturity, where income reliability is reduced.
Flags purchase options, superior lien claims, landlord covenant violations in co-tenancy leases, hazardous substance risks, nuisance uses, and other impermissible provisions.
Confirms presence and syntactical clarity of primary economic terms, required insurance clauses, indemnification language, subordination provisions, and other standard commercial lease requirements.
Evaluates tenant type and use against lender-specific policy requirements. Configurable without modification to the core decision engine.
Tests TI allowances and landlord work commitments against per-square-foot caps, ensuring concession packages remain within policy parameters.
Every determination is accompanied by a written audit-ready narrative citing the applicable servicing standard and the specific gates evaluated.
When the lease passes every sequential test, the system issues an automatic approval. No human review required. The borrower receives confirmation and a written decision record immediately.
Where the lease fails specific gates but the deficiencies are curable by lease modification or supplemental documentation, the system issues a conditional approval with enumerated conditions.
Where one or more gates cannot be satisfied, the system issues a disapproval with a specific identification of the failed tests and the factual basis for each failure. The decision is appealable.
The system logs failed tests, validates supplemental data, re-runs affected gates, and, where remaining failures fall within predefined override bands, routes the package to a human credit committee with side-by-side metrics and a system recommendation. Appeal outcomes continuously refine decision thresholds.
What are the Bills of Rights? They express, in terms more detailed than ever before, what objective reasonableness really is. A borrower has a right to a fast process that does not frustrate their ability to manage their property. A borrower has a right to a level of respect and deference that recognizes their status as a multimillion dollar contractual counterparty, and that increases as a lender's loan derisks. Those rights are subject, however, to a superior right of their lender to block decisions that truly increase default risk, the risk of loss, or that impair the loan as an asset. The appeal process and human supervision enable continuous, organic improvement.
Every disapproval or conditional approval triggers an appeal right. The process is automated where possible, and controlled human intervention ensures decision quality and drives improvements.
Borrower or servicer submits an appeal request within the permitted window, uploading revised lease terms and supplemental documentation.
System verifies the loan and lease are eligible for appeal and that the submission is within the permitted period. Ineligible or untimely appeals are rejected automatically.
System logs the specific algorithm gates that caused the original disapproval, DSCR/LTV tests, prohibited provision tests, policy tests, or TI caps, and validates the supplemental data.
Only the affected gates are re-run against the new data. The system generates new DSCR, LTV, policy, and TI evaluations and determines whether all tests now pass.
Where tests remain failed, the system evaluates whether the shortfalls fall within predefined override bands, modest DSCR tolerance, minor LTV excess, limited policy deviation. Outside those bands, the appeal is denied.
Failures within override bands are routed to a credit officer or committee with side-by-side pre- and post-appeal metrics and the system's recommendation. The committee may uphold disapproval, approve with conditions, or approve without conditions. All outcomes are locked for audit, and successful appeals trigger algorithm refinement.
Approval Screen integrates with existing servicing platforms via API or deploys as a standalone secure web application. Contact us to schedule a demonstration.
A provisional patent application has been filed.