Automated Credit-Loan Platform based on AI-technology A fully - - PowerPoint PPT Presentation

Automated Credit-Loan Platform based on AI-technology

The Corex.ai system is a fully automated loan management system. This solution extends the productivity of Your loan business to the maximum. Easily integrates with other data service providers. From Data input to vali- dation, Credit History, Scoring & Rating to Decision making. Fully automated for the maximum security.

A fully automated processes

Data Input

Website registration/mobile application Through other affiliates and agents Agent API

Validation

Compliance checks Rule based checks

A fully automated processes

Credit History

Connectivity to third data providers Automated checks in internal loan databases

Scoring & Rating

Internal third party data rating system Personalized, flexible and internal loan data scoring

Decision

Manual/automated loan confirmations and rejections Automatic money transfer Possibility to rate the scores

End-to-end lending process

Corex.ai is an end-to-end integrated solution. This solution covers the full lending cycle from application, through loan origination, risk assessment, decision-making , monitoring and management control of loans. It can all be made to be automated or manually controlled according to your needs and your budget.

> Data input > Product setup (customer loans, payday loans, installment loans, secured loans, syndicated loans) > Pricing control > Marketing and promotion setup > Loan proposal > Automated/manual deduction > Late, penalty, tax calculator > External/internal bailaff control > Customer verification > Credit history check > Scoring and rating > Automated loan approval > Manual, semi-automated, automated payment process > Automated email and SMS/Notifications

E v a l u a t i

• n

a n d a p p r

• v

a l S e r v i c i n g a n d r e p

• r

t i n g O r i g i n a t i

• n

C

• l

l e c t i

• n

Lending life cycle

End-to-end lending process

Replacing manual steps with automated decision rules, transforming paper into digital format, and gathering performance data on all aspects of your lending practice lets you: Provide a better user experience to improve the speed and accuracy of loan applications Process a larger number of loan applications with existing resources Eliminate sources of human error for faster, better-quality evaluation decisions Establish predictable, repeatable, and auditable processes that support compliance Reduce delays and costs associated with paper processes Analyze process and loan performance with the goal of continually improving efficiency and profitability

Accurate and Repeatable Decisioning

Automation improves applicant evaluation and loan decisioning. It eliminates repetitive manual steps that are best executed digitally and allows human expertise to be applied where it works best. Examples include:. Decision rules evaluate applicant attributes and provide immediate loan approval for well-qualified applicants

• r declination for applicants who have no credit history, are underage, or over-extended credit.

Integration with credit data sources and services such as Corex AI Algorithms lets lenders automatically and quickly verify applicant information. Decision rule flexibility can guide processing for exception applications, performing certain steps of the evaluation process automatically and delegating more complex decisions to experienced underwriters. No lost or misplaced documents. Paper documents converted to digital images are immediately and securely accessible by the underwriter, so they can review applicant materials more quickly. The combination of decision rules and integration with credit data sources and services lets lenders automatically calculate optimum loan structures and terms. Decision rules provide predictable, repeatable processes that remove the variations associated with human decisions and demonstrate compliance with lending regulations.

Multichannel for Digital Notification

Automation decreases the manual work associated with notifi-

• cations. Every application requires a response, and a digital

economy gives lenders great flexibility in how to respond. An au- tomated loan processing system helps lenders create a variety

• f notification templates that are then populated with applicant

and loan-specific information before sending the notification. Corex platform communication gives lenders the ability to send notifications by email, print, and even text messages. One clear advantage of Corex platform is the reduction of print- ing and mailing costs. Even better: Secure, digital retention of no- tifications as records. Not only is this beneficial in the event of demonstrating regulatory compliance, as in the case of adverse action letters, but it also relieves lenders of the cost and risks in storing paper documents.

Analytics for Continuous Process Improvement

In addition to the advantages mentioned above, automation opens the door to analysis of your lending practices. An auto- mated process creates a wealth of application, underwriting, and funding data. Analytical tools specifically designed for lending allow you to slice and dice this data to determine where improvements can be made to current lending practices. Your analytic results will show you where you are already performing well (no need for improvement) and more importantly identify areas for lending practice improvement. With analytics tools you can: Identify the geographies that send you the best loan opportunities and determine how you might replicate this success in

• ther geographies.

Determine which borrower factors have the greatest influence on loan performance and use this insight to accept more applicants with these attributes. Discover which loan structures and terms reduce risk default for subprime borrowers. Identify borrower characteristics that lead to unexpected delinquencies and implement more stringent qualification steps to reduce risk. Identify unnecessary, time-consuming manual steps in loan origination that can be improved through the use of decision rules.

AI Process NL

Automated 89% of a manual process

A member submit a policy compliant The provides makes a decision

• n the compliant

The decision and date is recorded on the policy Automation Rate Accuracy Rate Improved accuracy of time stamp and other data extracted from

62% to 85%

Automated 89% of the manual process for extracting a date from the policy decision

The results

89% 85%

QCR PRA BPM

Digitalization policy with the compliant and recorded date Rules to extract data from the policy Manage process expertions

As online banking has allowed consumers to shop around for preferable rates, industry leaders are attempting to stymie this burgeoning competition by increasing cross-sell

• pportunities. Making experiences easy, seamless and

personalized should be a priority in any industry, but consumers increasingly expect it from financial services firms.

Customer experience in financial services

Onmichannel capabilities being implemented by financial firms in 2017

Customer relationship management Mobile banking Online account management Online lending Interactive ATMs % of banking executive survey respondents 0% 10% 20% 30% 40% 50% 60%

57% 47% 46% 31% 26% Trend attributed key:

Very Low Low Medium Hight Very Hight Maturity Amplitude Velocity Impact

Supervised learning-checking accuracy

Training Model Training Validation Validation Compare results calculate sccuracy score Predicted

• utput

Known correct

• utput

Prediction Known correct

• utput

Test set Training set 60% 40%

Trusted Personal data storaging

DocChain: Permissioned Blockchain

Identity Owner

SDK

Identity Provider

SDK

Identity Requester

SDK Data Store 1 2 Cold servers Hash Identity by Consensus

Identity Owner Identity Provider

Private Key

Trusted Key Holder Trusted Key Holder Trusted Key Holder

Part Key Part Key Part Key

Project Specification

Project Specification

Project Specification

Project Specification

Loss of profit due to incapability to evaluate a deal Human predictions have low accuracy - only 48% Credit specialists are afraid of decision-making Credit analysis managers operate with lots of mistakes Analytic team requires lots of expenses

Top-5 Challenges Faced by Market

Cloud Computing Solution Modular Architecture

Allows to unfold a system anywhere in the world immediately. ndependent from enterprise computing power. Allows to enrich the system with additional functions, if needed, without capital expenses.

White Label AI

Design was developed in a way that allows our specialists to customize system interface according to your corporate identity We developed unique algorithms that allow to normalize incoming data, process and analyze it with our systems. They proceed fast and adjust to new data. Thereby you get the most honest evaluation of each deal, considering previous experience of each counterparty collaboration

System Tasks

Predictive algorithms, utilized by employees, help to reduce the influence of ‘human factors’ on decision making and risk assessment Process automation, algorithmic assistance in business processes. Actual decision is taken by AI

• algorithms. Your employees

just have to validate and correct it (if needed). It takes 60% less time to process a transaction Analytical tools are flexibly tuned Simplifies work of a transaction support manager in a sales funnel from the moment of qualifying till payment receipt.

System Effectiveness in Q4 of 2018

In the 4th quarter of 2018 by using our algorithms our partners achieved significant results:

By using special mechanisms for normalization and risk assessment total revenue increased to a record

$105 million USD from $80 million USD

Due to preliminary forecast on each deal our partners could reduce by 28% losses from unscrupulous counterparties. Proved economy is

$15 million USD in total

System Effectiveness in Q4 of 2018

Average accuracy of our predictions is 78 - 86%. Comparing with human predictions, accuracy increases significantly: is a change of

45-55%

We predict for 2019 profit increase for our partners to the tune of $300 million

USD

Technology

To evaluate each operation, system uses more than 100 specially trained neural networks. It allows to consider all factors, even ‘insignificant’ Improving risks of individual counterparties. Algorithm uses technology called ‘computer vision’ to check counterparty’s psychological type and to reveal person’s unscrupulous tendencies Precise normalization of incoming data, preliminary assessment and data supplement from additional sources (open databases, corporate and financial reports, declarations, social networks). Machine learning technologies research a million times faster than humans Do we have competitors? - Yes, we do. What makes us different? - We are professionals in platform automation. We make tech solutions for our partners that not only adapt to their inner business processes easily, but also from the very first moment bring concrete results with each counterparty.

Advantages

Cloud solution, fast to unfold, independent from partner’s computing power Accelerates processes, eliminates ‘human error’ Considers even indirect features that influence on business results of each operation All components of the system can infiltrate independently, but at the same time are compatible with a third-party software Report can be preset or put together individually We can do either way, so it’s up to you! Flexible customization. Exactly what you need for your inner business processes Innovative methods of operational evaluation Outcome since first minutes

Retail &Wholesale Correspondent Automated Processing Expert Underwritting Closing / Funding

Multi-Channel LOS Portals

Relationship Management Contact, Appointment Leads, Loan Scenarios Metrics, Pre-Approvals 1003, Loan Status

Consumer Direct

Quick Qualifer Quote Personalized Rate Qoute Online Loan Application Upload Docs, eSign Auto Income & Assets

Mobile Applications

Quote & Apply Online Consumer Mobile Portal Loan Officer Mobile Wholesale Mobile AE’s, Brokers, Realtors

Reusable Lending Services

form the basis of a microservices architecture promoting robust consistent results across all applications

Fully Managed Modern Platform

highly scalable, cloud deployed, thin-server business and data layers

Product Eligibility Loan Conditions Pricing Engine Decision Service Morgage Calcs Fees Engine Quote Service Checklist Service Data Calculations

Configuration

Lender specific policies and compliance regulations

Security

Platform end-to-end security, monitoring and performance encrypted by key Blockchain

Automation

Process automation driven by rules engine inteligence, ML, NLP

Amazon API or Azure API

Business and data domain thin-server based web services

Data model Synchronous Replication Reporting Read-Replica based on Blockchain

Robust flexible consumer centric design

How it works

OUR SYSTEM

Feed FTP API

Risk scoring Result prediction Individuality scoring Deal optimization recomendations Reports scheduler Data visual studio

API API API Lead data BI8

Lead data Deal data Deal/Lead management Predictions Risk management

API’s Feeds

• Ext. DB’s

Fin reports Social networks

Company clients

API’s

Marketing Contacting tools SMS Messengers Emails Call center Push notifications Flexible reporting engine

API

Some of our projects

Contact Details

address

Kyiv, Vozdvyzhenska 36

phone

+380 67 223 2960

e-mail

sales@corex.ai