Directory of AI consumer credit scoring and underwriting vendors

AIZEN Global runs two connected businesses from Seoul. ABACUS is an automated machine learning platform built for finance, used by large banks, card issuers and insurers to build, monitor and update thousands of predictive models in parallel through one interface, covering underwriting, fraud detection, anti money laundering, auditing and claims, and exposed through interfaces that leave existing systems unchanged. CreditConnect applies it to lending, converting non financial data from e-commerce, mobility, e-wallet, education and healthcare platforms into credit decisions so those platforms can offer financing while banks remain the lender. The company was the first designated by Korea's Financial Services Commission to underwrite loans on behalf of banks using AI-driven credit decisions, and signed 117 affiliated companies within eight months of launching in Vietnam.

Singapore company founded 2019 by Hayk Hakobyan selling alternative credit scoring and financial intelligence to banks, fintechs, e-commerce companies and telecoms across Southeast Asia and beyond, aimed at the unbanked and underbanked. Products include a Financial Health Profile for individuals and a Financial Business Health Profile for micro and small enterprises, alongside fraud detection, eKYC and a product recommendation engine. Risk profiles are built from financials, health, lifestyle and social footprints, and its own account of the method includes personality based and voice based risk assessment to determine loan suitability. HSBC's venture arm is an investor.

Carrington Labs builds cash flow underwriting models and credit risk analytics for banks and non-bank lenders across consumer and small business lending, as a business of an Australian listed group. Its position is that most lenders already have the data and the decision engine and the gap is the model in between, so it builds that model from the lender's own borrowers and repayment outcomes and delivers it into systems already running, explicitly not replacing origination, decisioning or servicing. Its Cashflow Score runs solely on customer-permissioned bank transaction data, returning a 1 to 100 value derived from five named categories of credit risk behaviour, backed by the performance of millions of loans, and reports up to 30 percent higher accuracy than traditional credit models. Coverage spans underwriting, loan and limit sizing, risk-based pricing, post-origination limit management and early warning detection, with model features mapping directly to adverse action reasons. It launched the first protocol server letting lenders call institution-specific credit model outputs inside agentic workflows.

CredoLab scores creditworthiness from smartphone device metadata for banks, consumer finance companies, auto lenders, online and mobile lenders, insurers and retailers, aimed at applicants with no credit file. A white labelled app or embedded kit collects behavioural signals only after explicit opt-in, covering application ownership patterns, device model and age, contact and message counts, file sizes and interaction habits, and the company states no personally identifying information leaves the device and that it never learns an applicant's name, address or number. Models built on more than 21 million loan applicants across 70 lending partners have supported over a billion dollars of lending in more than 20 countries, with behavioural patterns learned across 50. A 2025 income prediction model estimates earnings from thousands of anonymised signals, and institutions can train it on their own local populations.

CRIF is an Italian credit information and decisioning group founded in Bologna in 1988, operating credit bureaus, business information services, analytics, outsourcing and processing across roughly forty countries on four continents with more than six thousand six hundred professionals. It reports supporting over five thousand banks, financial institutions and leasing companies in end to end credit management, alongside tens of thousands of non financial businesses, and it also serves insurers, telecom and media operators and energy and utility companies. Its regulatory position is unusual for a technology supplier and rests on two distinct authorisations: CRIF Ratings is a credit rating agency registered with the European Securities and Markets Authority and recognised as an External Credit Assessment Institution, issuing ratings on non financial companies based in the European Union, and the group is an authorised Account Information Service Provider in every European country where the second payment services directive applies. Several central banks and public authorities use its technology to run national credit reporting infrastructure, and it has built bureaus in markets as varied as continental Europe and the Caribbean. The software line comprises a lending journey platform covering digital onboarding, identity and business verification, open banking data and creditworthiness assessment, and an end to end credit management platform that calculates scoring and rating models, drives portfolio strategies such as pre approved offers, and runs early warning processes on traditional bureau and current account data together with categorised open banking data. Machine learning and generative capability appears as automated checks that verify each transaction against internal credit policy and regulatory requirements, a credit agent that surfaces real time insights and recommendations for credit managers, and algorithmic creditworthiness scoring inside the origination flow.

Crisil is a Mumbai headquartered analytics group founded in 1987 as India's first credit rating agency, now majority owned by S&P Global. Its ratings arm is registered with the Securities and Exchange Board of India, and its solutions business, rebranded from Crisil Global Research and Risk Solutions to Crisil Integral IQ, sells research, risk, lending, analytics and operations capability to global financial institutions alongside a set of licensed software platforms. The flagship is Credit+, covering the corporate credit lifecycle from initiation and financial spreading through credit risk rating, portfolio monitoring, early warning signals and covenant tracking, with the early warning module reported as implemented at more than ten banks and built on rule based flags and multi dimensional trigger libraries. Adjacent platforms include a Capital Assessment Module for automating credit assessment reporting and risk calculation, Model Infinity as a cloud ready platform for model inventory, workflow and governance, and a Scenario Expansion Manager for defining and analysing regulatory and internal stress testing scenarios. Generative capability was launched in 2025 as Crisil GenEye Credit and Crisil DeepMine, alongside generative services embedded into trading, finance, risk and credit workflows. The company publishes measured performance for its own automation, stating 95 percent data extraction accuracy for financial spreading with efficiency gains of 30 to 50 percent, and generative coverage of 60 to 70 percent of credit report sections yielding more than 30 percent efficiency. Model validation is a distinct and long standing business line, with more than 25,000 models validated for clients since 2015. Chartis named the company a category leader in model validation for a fourth consecutive year, assessing its generative capabilities across model development, validation, governance, inventory management and risk control, and placed it in the RiskTech100 for a third consecutive year.

Independent risk technology firm operating since 2010, whose central product ECOLO scores the credit risk of small and medium businesses from real operational business metrics rather than from financial statements. The company states it uses proprietary machine learning models to quantify tangible business metrics and derive the true revenue and profit of a business, and positions the result as approving small business loans in seconds with no financials required. It describes itself as having spent its first six years exclusively on market analysis, product development and building the underlying competencies before selling. A second line, YMIX, is a consulting engagement rather than software: mathematical and logical reengineering of the calculations behind retail banking products, aimed at finding operational leakage so a bank earns marginally more per transaction. A third strand is published analytics, including a Small and Medium Enterprise Pulse index tracking the development of that segment across national economies, and research on the shadow economy. A borrower facing lending platform under the ECOLOplace name is described as in development. Named clients on its own site are predominantly banks across Central and Eastern Europe and Central Asia, including Intesa, OTP Bank, Societe Generale, UniCredit, Jusan, Hamkor and Fast Bank. Aggregate figures published are more than 100 percent return on investment per project, under 1 percent non performing loans on unsecured small business lending, more than 50 projects worldwide and 830 million dollars of additional income for customers, none of them attached to a named institution. Appears on the Chartis Credit Lending Operations 2026 roster and the Chartis Credit Risk Management 2025 roster.

FinbotsAI builds credit scorecards for lenders through creditX, a no code platform that lets a bank's own credit modellers generate and deploy high accuracy scorecards in hours rather than months, choosing their own data and parameters while machine learning improves the resulting model over traditional regression methods. It serves banks, digital banks, small business and consumer lenders, fintechs and credit bureaus across Asia, Australia, the Middle East and Africa, and its scoring solution has completed the Singapore government's artificial intelligence governance testing framework and the central bank's responsible AI programme for financial services.

Finvero runs a multi-lender credit marketplace in Mexico and Colombia connecting lenders, merchants and consumers, supplying the credit infrastructure and pre-qualified applicants rather than lending itself. Its four modules cover origination, a risk and fraud engine, collections and portfolio administration, across both consumer and business segments and product types including instalment, revolving and buy now pay later, with in-store origination. Alternative credit scoring built on generative AI and alternative data supports decisions in under five minutes, and the company reports lenders improving decision accuracy by 10 to 15 percent. Lenders build their own traditional, AI and predictive scoring models and set their own fraud criteria on the platform. Its collections model publishes its full feature set, and it partners with a card network's inclusive growth programme supporting micro-entrepreneurs.

GiniMachine is a no-code credit scoring platform that builds, validates and deploys machine learning risk models from a lender's own historical loan performance data in seconds to minutes, aimed squarely at institutions with no data science team. Built by a fintech product company and integrated with its lending suite, it uses decision tree methods with automated model construction, monitors its own models, and is positioned explicitly against black-box tools as a transparent web application with interface access. It scores applications using alternative data including rental and utility payments, asset ownership and public records to reach thin-file borrowers, lets the lender set its own cut-off and risk tolerance, and extends to collections by prioritising debtors likely to repay and suggesting the most effective contact method. Coverage spans online, commercial, point-of-sale, auto and card lending alongside small business finance, factoring and leasing.

Karus builds credit intelligence for consumer auto finance, serving originators, lenders, dealers and the investors who buy the paper, with proprietary models trained on tens of millions of loan outcomes. Its argument for why auto is its own discipline is that most loans move through a dealer rather than direct to the borrower, so an originator must price in seconds against the borrower's credit trajectory and the specific vehicle's depreciation risk: generic tools score the borrower, auto requires scoring the loan and the dealer. Separate model classes handle underwriting, loan structuring and dealer level pricing, with real time portfolio monitoring, cash flow forecasting and a conversational layer over the suite. The platform ran alongside a lender's own underwriting team for twenty months on the same loan population as a controlled comparison.

Monnai supplies consumer insight infrastructure to digital lenders, banks and fintechs across the United States, Latin America, India and Southeast Asia, delivering four decisioning modules through a single interface: customer identification, trust and fraud risk, credit decisioning and collections optimisation. It aggregates, normalises and contextualises disparate data sources across silos and borders, drawing on payment, communication, device and identity signals, and enriches coverage with proprietary data for geographies where verification is otherwise hard. The company states it can return hundreds of insights on billions of consumers through one interface, and reports customers seeing 99 percent detection of fraudulent identities alongside a 40 percent increase in approval rates and a 45 percent reduction in defaults. A graph based dashboard lets fraud and credit analysts identify risk factors in a single view.

Moody's Analytics is the data, research and decision solutions division of Moody's Corporation, scoped here at division level and separated from Moody's Ratings, a boundary the company itself draws on every product page. It sells to banking, buy side, insurance, corporate and public sector customers across ten solution lines including lending, third party risk, regulatory reporting, balance sheet management, insurance underwriting, portfolio management and model risk governance. The shipped inference line has two generations. Research Assistant launched in December 2023 as a conversational interface built on Microsoft Azure OpenAI over the proprietary data estate, covering more than 190,000 companies. Agentic Solutions followed in late 2025 as coordinated specialised agents across five workflows: company credit assessment producing a complete credit memo, portfolio monitoring with early warning signals, sales intelligence, customer and counterparty screening covering sanctions and adverse media, and private credit risk assessment. Both run over a context layer spanning more than 600 million entities. The Lending Suite absorbed Numerated Growth Technologies, acquired in November 2024, adding a loan origination system used by institutions holding a combined 3 trillion dollars in assets that had processed over 65 billion dollars in lending. Named AI partners include Anthropic, OpenAI, Microsoft, Databricks, Amazon Web Services and Salesforce.

Omnisient runs a privacy preserving data collaboration platform that lets banks, insurers and credit bureaus draw alternative data insights from retailers, telecommunications operators and other consumer businesses without either side exchanging personal information. Privacy enhancing technologies, tokenisation and cryptography hold the parties apart while embedded machine learning tools build and test scoring models inside a neutral environment, so only insights move rather than data. The principal use case is credit risk scoring for consumers with no credit history or thin files, alongside fraud and financial crime work and payment media network monetisation for banks.

Optasia is a listed AI credit decisioning platform embedded inside mobile operator and wallet ecosystems across 38 countries in Africa, the Middle East and South Asia, working through 49 distribution partners and 13 banks. More than 200 machine learning models draw on thousands of alternative signals per user to make around 1.5 billion credit decisions a month at roughly 300 per second, supporting micro loans averaging about five dollars and airtime advances that traditional banks cannot profitably process. It does not lend from its own balance sheet: partner banks provide liquidity while Optasia underwrites the default risk and provides guarantees, letting distributors earn lending revenue without balance sheet exposure. Built specifically for markets, it facilitated around six billion dollars of credit in a year, serves over 430 million annual active users, and listed on the Johannesburg exchange in November 2025.

Orbii supplies credit infrastructure to digital lenders, fintechs, payment companies and banks across the Middle East so they can launch and run small business lending without building credit functions themselves. It connects directly into the systems businesses already operate, including point of sale terminals, enterprise resource planning software and banking channels, collects borrower financial data automatically through interfaces or statement uploads, enhances that raw data into a view of financial health, and runs machine learning models that underwrite, disburse and monitor loans in real time. Its argument is that conventional credit assessment fails in markets with thin bureau coverage and fragmented data, so reading actual trading activity produces both higher approval rates and lower defaults. It has processed thousands of applications and targets a billion dollars of small business lending.

Pave turns a lender's own raw bank transaction data, credit reports and loan performance history into cashflow based credit signals, producing more than 4,000 attributes across affordability, stability, willingness and assets alongside scores trained separately for each credit product and, for small business lending, for each industry. The stated purpose is finding creditworthy borrowers a bureau score alone would miss, and the platform is used for underwriting, lead scoring before a credit pull, dynamic credit limit setting and prioritising collections by who currently has cash to pay. The company is explicit that it supplies no data of its own and that its analytics enhance a lender's proprietary models rather than replacing them. It operates at pavefi.com, formerly pave.dev.

Prism Data pioneered cash flow underwriting, turning consumer bank account transaction history into a three digit score lenders can drop into existing credit policies alongside a bureau score. Its CashScore is consortium based, built from millions of anonymised, consumer permissioned records spanning many banks, credit products and customer segments, and is offered alongside first party fraud and small dollar lending variants plus income, categorisation and trended attribute products. Data reaches it de identified through any aggregator, decision engine or single endpoint interface, and returns in under a second. The company states compliance with United States credit reporting and equal opportunity law, supplies adverse action reason codes, and offers delivery through both consumer reporting agency and non agency channels.

Scienaptic AI provides credit decisioning to US credit unions, banks and lenders, building scorecards on each client's own loan book augmented by more than 3,000 signals across bureau, banking and alternative data, and reporting twelve times more risk differentiation than bureau scores alone. It reports approving up to 40 percent more members, raising approval rates for protected classes by more than 45 percent, and assessing over 90 percent of individuals without traditional credit histories, with 60 to 80 percent of decisions automated and fair lending monitoring built into the platform. Synthetic identity, bot attack, bust-out, credit washing and first-party fraud are flagged inside the same decisioning call before underwriting sees the application. Its current product adds language models and agentic capability, framed as putting humans back at the centre of lending, and it integrates natively into a major core banking provider's loan origination system.

Stratyfy builds interpretable machine learning for financial institutions across credit risk assessment, fraud detection and bias mitigation, on the argument that transparency and control matter more than raw predictive power when the decision affects a person. Its Probabilistic Rules Engine produces decisions expressed as readable rules rather than scores, so any prediction can be explained directly to the customer, to regulators and to internal stakeholders, and lenders can write their own knowledge of market conditions and emerging risks into the model alongside the data. A published comparison against conventional decisioning found it identified nearly twice as many pre qualified applicants while lowering the overall bad rate, expanding the addressable market by 70 percent. The company is women led and backed by a major bank's venture arm and a core banking provider.

Traive supplies credit risk assessment and asset qualification across the agricultural finance chain, serving lenders, crop input manufacturers, traders, cooperatives and capital markets participants rather than farmers directly. Its models combine language models with generative adversarial networks to evaluate more than 2,500 data points per borrower, drawing on alternative farming data and individual producer behaviour, and a real time monitoring system tracks exposures using analytics connected to satellite imagery. A single platform lets participants review, register, manage and trade agricultural loans, with the stated aim of turning farm credit into liquid, qualified financial assets that capital markets can hold. Founded in Brazil with a United States base, its Series B was led by the largest agricultural lender in its home market.

Trusting Social scores consumers with little or no formal credit history for more than 130 financial institutions across Vietnam, Indonesia, India and the Philippines, using proprietary machine learning over alternative social, web and mobile data. It reports having scored over a billion consumers, counts more than 40 institutional clients in Vietnam and six of the ten largest banks in the Philippines, and extends beyond scoring into digital identity verification, fraud prevention and an unusual model that predicts residential and office addresses from alternative data. The company also builds co-lending and embedded finance arrangements with banks and consumer brands, including a platform partnership with a Vietnamese conglomerate aimed at reaching 27 million families. Founded in 2013 by a data scientist with a doctorate in econometrics and a background in global credit risk at a major bank.

Upstart operates an artificial intelligence lending marketplace through which more than 100 banks and credit unions use its underwriting models and cloud applications to originate consumer credit. Its personal loan model weighs more than 3,000 variables and retrains against loan level repayment and delinquency data arriving daily across the whole partner base, covering personal loans, automotive retail and refinance lending, home equity lines and small dollar relief loans. Lending partners set their own credit policy, objectives and risk appetite and remain the lender of record. The company publishes annual comparisons of its model against a traditional benchmark including outcomes broken out by borrower race and ethnicity, maintains hundreds of state licences, and has applied for a national bank charter to be held in a separately regulated sister company.

Zest AI has built machine learning credit underwriting for US lenders since 2009, serving institutions from the largest banks and auto and specialty lenders down to credit unions processing as few as a hundred applications a year. Its distinguishing capability is fairness engineering rather than accuracy alone: its technology searches for less discriminatory alternatives, the legal standard under US fair lending law, and applies adversarial debiasing to reduce disparity identified during model fair lending testing. A model management system lets credit teams build, validate, deploy and monitor their own underwriting models, so the lender owns and controls the model rather than outsourcing decisions to a marketplace. Alongside underwriting it offers application fraud detection and generative insights drawn from industry and macroeconomic data. With a credit union partner it created a cooperative service organisation so small institutions can access the same technology.