Concepts and terminology

Term	Meaning
AIOps	The Cortex data-science stage that derives analytical features and risk signals from normalized applicant data.
Actionable Intelligence	The final Cortex engine, which combines enriched data, risk configuration, and rules into an explained recommendation and onward actions.
BVN	Bank Verification Number, a unique banking identity number used in Nigeria.
Capacity to pay	An assessment of repayment ability using cash flow, affordability, and available security.
CIM	Common Information Model, the canonical schema into which applicant data is normalized.
Config plane	Merchant Activation: the tenant-specific policy layer that governs downstream enrichment and decisioning.
Credit Risk Framework	The scoring categories and point budgets that define how the TrustScore is structured.
CRIPE	Credit Risk Intelligence Platform for Enterprise, the unified Cortex platform hosting the five-engine underwriting flow.
Direct debit	A consent-based instruction that permits an authorized amount to be collected from a bank account.
DBR / IIR / DSCR	Debt Burden Ratio / Income-to-Installment Ratio / Debt Service Coverage Ratio, measures that compare repayment obligations with income or cash available for debt service.
DTI / EDTI	Debt-to-Income / Effective Debt-to-Income, measures comparing debt obligations with income.
EDA Store	Exploratory Data Analysis Store, where low-confidence or unparsed data can be held for review.
ETL	Extract, Transform, Load: the stage that converts source inputs into the normalized model used downstream.
GSM	Global Standing Mandate, a consent-based multi-bank direct-debit mechanism used by Recova Pro.
Ingestion	The secure gateway that associates incoming data with its tenant and source provenance.
Origination	The intake stage that captures applications from supported channels and sources.
Override trail	The record showing whether an authorized person changed a computed outcome and why.
Propensity to pay	A forward-looking assessment of repayment likelihood based on available behavioral and financial signals.
RADAR	CreditChek account-discovery capability used to identify bank accounts linked to a borrower's BVN.
Recova Pro	CreditChek's mandate and automated collection service for consent-based repayment.
Risk modelling	The weighted factors that translate risk signals into score contributions.
Risk signal	A derived indicator—such as income stability, affordability, delinquency, or exposure—used in assessment.
Rules	Hard policy conditions that can enforce an outcome independently of the weighted score.
Spectrum	CreditChek infrastructure for reporting loan-lifecycle activity to licensed credit-bureau partners.
TrustScore	CreditChek's explainable, points-based credit score with category and reason-level attribution.
Willingness to pay	A behavioral assessment of repayment intent based on available credit history and repayment patterns.

Decision engine response

The decision engine response combines the final recommendation with the evidence used to reach it. Field names and string values retain their response casing. A nullable field can return null when the value is unavailable, is not applicable, or cannot be calculated from the available data.

Top-level fields

Key	Type	Meaning
metadata	object	Identifiers that connect the response to the assessment, business, and applicant.
decision	object	The engine's final lending recommendation, amount, confidence, and decision timestamps.
trust_score	object	The overall TrustScore result and its explainable scoring breakdown.
willingness_to_pay	object	Assessment of the applicant's demonstrated repayment behavior and intent.
capacity_to_pay	object	Assessment of whether the applicant has sufficient, verifiable resources to repay.
risk_factors	array of objects	Material risks identified during assessment, with an explanation and severity for each one.
macro_context_summary	string	Summary of relevant macroeconomic conditions considered alongside applicant-level evidence.
explanation	string	Plain-language explanation of the main evidence and rules behind the final recommendation.
override	object	Indicates whether the computed TrustScore outcome was overridden and records the reason.

Metadata

Key	Type	Meaning
metadata.assessment_id	string	Unique identifier for this assessment run. Use it to trace the request, response, and related records.
metadata.business_id	string	Unique identifier for the merchant or business account under which the assessment was performed.
metadata.bvn	string	Applicant's Bank Verification Number. Treat this value as sensitive identity data.

Decision

Key	Type	Meaning
decision.recommendation	string	Final action proposed by the engine, such as approve or decline. A recommendation can reflect both scoring and hard policy rules.
decision.approved_amount	number	Maximum principal amount approved by the engine in the product's configured currency. A value of 0.0 means no amount was approved.
decision.confidence	string	Categorical indication of how strongly the available evidence supports the recommendation, such as low, medium, or high; it is not a probability.
decision.conditions	object	Proposed commercial terms, repayment structure, covenants, and special requirements attached to the recommendation. Conditions may be present even when individual terms are not yet available.
decision.decided_at	datetime string	Time at which the engine produced the decision, including its UTC offset.
decision.saved_at	datetime string	Time at which the decision was persisted, including its UTC offset. This can be later than decided_at.

Decision conditions

Key	Type	Meaning
decision.conditions.interest_rate	number or null	Interest rate proposed for the facility, using the rate convention configured for the lending product. null means no rate was produced.
decision.conditions.tenor_days	integer or null	Proposed loan duration in calendar days. null means no tenor was produced.
decision.conditions.repayment_frequency	string or null	Proposed repayment interval, such as daily, weekly, or monthly. null means no frequency was produced.
decision.conditions.covenants	array of strings	Ongoing obligations or restrictions the borrower must observe during the facility. An empty array means no covenants were returned.
decision.conditions.special_conditions	array of strings	Additional requirements, exceptions, or decision-specific conditions that must be noted or satisfied. An empty array means none were returned.

TrustScore

Key	Type	Meaning
trust_score.score	number	Final TrustScore on the platform's reporting scale after the scoring model has processed the available factors.
trust_score.recommendation	string	Score-level action indicated by the TrustScore. Its uppercase presentation, such as DECLINE, does not change its meaning.
trust_score.breakdown	object	Detailed attribution showing how the TrustScore was constructed and interpreted.
trust_score.breakdown.trust_score	number	The final TrustScore repeated inside the breakdown so the detailed attribution is self-contained.
trust_score.breakdown.risk_band	string	Human-readable risk tier mapped from the final TrustScore, such as High Risk.
trust_score.breakdown.raw_achieved	number	Total unscaled points earned from the applicable scoring factors before conversion to the reported TrustScore scale.
trust_score.breakdown.raw_max	number	Maximum raw points available under the configured scoring framework. It provides context for raw_achieved.
trust_score.breakdown.category_breakdown	object	Raw points earned within each configured scoring category.
trust_score.breakdown.category_breakdown.Personal Details	number	Points contributed by applicant-profile factors such as age, education, marital status, dependants, and residence.
trust_score.breakdown.category_breakdown.Policy Norms	number	Points contributed by policy-defined eligibility and operating factors, such as employment or business characteristics.
trust_score.breakdown.category_breakdown.Financial Analysis	number	Points contributed by affordability, income, turnover, and other financial-performance factors.
trust_score.breakdown.category_breakdown.Credit History	number	Points contributed by bureau history, active facilities, enquiries, repayment performance, and related credit behavior.
trust_score.breakdown.reasons	array of strings	Factor-level audit trail showing the points earned or why no points were awarded. For example, +50 is a positive contribution, while 0 (no data or poor indicator) means the factor earned no points because evidence was missing or did not meet the configured threshold.

The affordability reason DBR / IIR / DSCR groups three related measures: Debt Burden Ratio compares recurring debt payments with income, Income-to-Installment Ratio compares income with installment obligations, and Debt Service Coverage Ratio compares cash available for debt service with required debt payments. The exact thresholds and weighting depend on the configured credit policy.

TrustScore reason labels

Label	Meaning
Age	Points based on the applicant's age and the configured eligibility or risk ranges.
Educational Qualification	Points based on the highest verified education level supplied for the applicant.
Marital Status	Points assigned to the applicant's declared marital-status category under the configured model.
Dependents	Assessment of the number of people financially dependent on the applicant and its effect on disposable income.
Home Ownership	Assessment of the applicant's housing tenure, such as owned, mortgaged, or rented.
How Long Have You Lived Here	Length of residence at the current address, used as a stability indicator.
Age in Employment	Length of time in the current employment or occupation, used as an income-stability indicator.
Business Location	Assessment of the business's operating location using the configured location and verification criteria.
Business Registration	Whether the business has a recognized and verified registration.
Net Income	Income remaining after applicable expenses or deductions, used to assess resources available for repayment.
DBR / IIR / DSCR	Combined affordability assessment based on debt burden, income relative to installments, and cash available for debt service.
Banking Turnover	Value of account activity over the analysis period, used as evidence of business or income volume.
Outward Cheque Returns	Cheques issued from the applicant's account that were returned unpaid; repeated returns can indicate liquidity or account-management risk.
Repayment of Existing Active Loans	Observed payment performance on loans that are currently open or being serviced.
No. of Enquiries	Number of recent credit-information searches recorded for the applicant. A high count can indicate frequent credit seeking.
No. of Loans Servicing	Number of active loans on which the applicant is currently making or expected to make payments.
Overall Score	Any configured composite or upstream score used as an additional scoring factor. The reason output shows whether it contributed points.

Willingness to pay

Key	Type	Meaning
willingness_to_pay.score	number	Component score for repayment behavior and intent on the configured willingness-to-pay scale.
willingness_to_pay.classification	string	Risk label derived from the willingness-to-pay evidence, such as High Risk.
willingness_to_pay.positive_signals	array of strings	Favorable repayment-behavior evidence identified by the engine. An empty array means no positive signals were identified from the available data.
willingness_to_pay.risk_signals	array of strings	Adverse behavioral or credit-history indicators that weaken confidence in repayment intent.
willingness_to_pay.behavioural_observations	array of strings	Plain-language findings derived from the applicant's observed repayment and bureau history.
willingness_to_pay.summary	string	Concise narrative explaining the willingness-to-pay assessment.
willingness_to_pay.verdict	string	Overall acceptability judgment for repayment willingness, such as Acceptable or Unacceptable.
willingness_to_pay.recommended_lending_confidence	string	Suggested confidence level for lending based specifically on willingness-to-pay evidence. This can differ from the confidence on the final combined decision.

Capacity to pay

Key	Type	Meaning
capacity_to_pay.score	number	Component score for affordability and repayment ability on the configured capacity-to-pay scale.
capacity_to_pay.classification	string	Capacity category derived from available financial evidence. Insufficient Data means the engine could not verify enough evidence to make a reliable capacity assessment.
capacity_to_pay.recommended_monthly_repayment	number or null	Estimated monthly repayment the applicant can support. null means a reliable repayment amount could not be calculated.
capacity_to_pay.recommended_loan_amount	number or null	Estimated principal amount supportable by the assessed capacity. null means a reliable loan amount could not be calculated.
capacity_to_pay.positive_signals	array of strings	Favorable income, cash-flow, employment, or affordability evidence. An empty array means none was identified from the available data.
capacity_to_pay.collateral_valuation	object	Assessment of security or receivables available to support the proposed exposure.
capacity_to_pay.cash_flow_metrics	object	Calculated measures describing account inflows, their stability, and the relationship between inflows and outflows.
capacity_to_pay.risk_signals	array of strings	Adverse or missing financial evidence that reduces confidence in repayment capacity.
capacity_to_pay.cashflow_observations	array of strings	Plain-language findings derived from cash-flow analysis. The response uses cashflow without an underscore in this key.
capacity_to_pay.summary	string	Concise narrative explaining the capacity-to-pay assessment.
capacity_to_pay.verdict	string	Overall sufficiency judgment for repayment capacity, such as Sufficient or Insufficient.
capacity_to_pay.lending_recommendation	string	Lending action suggested by the capacity analysis alone. The final decision can also incorporate willingness, TrustScore, policy rules, and other risks.

Collateral valuation

Key	Type	Meaning
capacity_to_pay.collateral_valuation.collateral_type	string	Form of security supporting the exposure. Unsecured means no pledged collateral supports the facility.
capacity_to_pay.collateral_valuation.receivables_backed	boolean	Whether eligible invoices or other receivables are being used as repayment support or security.
capacity_to_pay.collateral_valuation.total_invoice_value	number or null	Combined face value of invoices submitted for receivables-backed assessment.
capacity_to_pay.collateral_valuation.advance_rate_percentage	number or null	Percentage of eligible receivables that may be advanced as financing, before other limits or deductions.
capacity_to_pay.collateral_valuation.discount_fee_percentage	number or null	Percentage fee or discount applied to receivables financing.
capacity_to_pay.collateral_valuation.eligible_invoices_count	integer or null	Number of invoices that satisfy the configured eligibility rules.
capacity_to_pay.collateral_valuation.unsecured_portion	number or null	Part of the proposed exposure not covered by recognized collateral or eligible receivables.
capacity_to_pay.collateral_valuation.summary	string	Plain-language summary of the security assessment.
capacity_to_pay.collateral_valuation.verdict	string	Overall collateral judgment. None indicates that no collateral was provided or assessed, rather than a positive or negative collateral quality rating.

Cash-flow metrics

Key	Type	Meaning
capacity_to_pay.cash_flow_metrics.monthly_average_inflow	number	Average incoming account value per month over the analysis period, in the account's currency. High inflow alone does not prove that income is stable or available for repayment.
capacity_to_pay.cash_flow_metrics.income_stability	string	Categorical assessment of how consistent verified inflows are over time, such as STABLE.
capacity_to_pay.cash_flow_metrics.inflow_outflow_ratio	number or null	Ratio of total or average inflows to outflows over the analysis period. A value above 1 generally means inflows exceed outflows; null means it could not be calculated reliably.

Risk factors

Each entry in risk_factors uses the following fields:

Key	Type	Meaning
factor	string	Short name of the risk area, such as Credit History, TrustScore, or Graph Risk. Graph risk refers to risk inferred from relationship or network checks; an unavailable graph check may cause the engine to apply a conservative buffer.
detail	string	Applicant-specific explanation of the evidence or condition that created the risk factor.
severity	string	Relative materiality of the risk, commonly low, medium, or high. Severity helps prioritize factors and is not itself a probability of default.

Context and explanation

Key	Type	Meaning
macro_context_summary	string	Narrative of relevant external conditions—such as inflation, monetary policy, interest rates, or default pressure—that may affect affordability and portfolio risk. Macroeconomic context supplements rather than replaces applicant-level evidence.
explanation	string	Consolidated reason for the final recommendation, including decisive rules, critical alerts, missing evidence, and the strongest risk findings.

In macroeconomic summaries, CBN means the Central Bank of Nigeria and MPR means Monetary Policy Rate, the benchmark policy rate set by the CBN. Delinquency means a repayment remained unpaid after its due date; overdue days measures how long it remained past due. A critical alert is a signal configured to trigger or strongly influence a policy outcome, including an automatic decline where the applicable rule requires it.

Override

Key	Type	Meaning
override.agent_overrode_trustscore	boolean	Whether an authorized decision agent or reviewer replaced the outcome implied by the computed TrustScore. false means no override occurred.
override.override_reason	string or null	Recorded justification for an override. It is normally null when agent_overrode_trustscore is false.