Linking AI Initiatives to Real-World Outcomes

Use Case

BANKS & FINANCE ORGANIZATIONS

Client

Banks/finance organizations. Most financial software was designed and developed prior to the mass selling of financial products on the Internet. Many struggle to meet the needs of an automated, straight-through-processing (STP), AI driven world.

Problem

Most banks/finance organizations are restricted by their platform, and some struggling to offer and automate digital products. Growth and profitability are constrained. Time to market, adapt and innovate are too long. Using traditional, pre digital product platforms or trying to adapt them is not working efficiently.

Banks and insurance companies moving their products and services on-line or via an app; they are automating many, previously manual processes. The rate of this “digital transformation” is accelerating.

For technology this brings challenges:

  1. Transforming a product from off-line to digital is not easy.

  2. Most systems and software tools were not designed to do this.

  3. Trying to adapt them is time consuming and expensive, and the results are usually sub-optimal.

  4. A large range of new functionality, data, context, knowledge and analytics is required to manage scalable, digital finance.

The issue is just as with selling and servicing products “face to face”, to sell on-line effectively you need to “see” and respond to your customer individually. Additionally, this response needs to be intelligent, high speed and in real-time. This requires a whole new range of predictive data and artificially intelligent, machine learning systems. This is not easy if your system was not designed to do this from the outset.

Implementation Approach

A survey of banking/finance companies in the process of moving to digital products was carried out. They were asked about the limitations of their existing system and problems they encountered. Here are some of the responses.

UI

  • Lack of personalization capabilities
  • Decision workflow hard to customize to match digital product requirements
  • Inflexible customer journey and adaptation
  • User interface changes take too long and are expensive to implement, limiting innovation and experimentation

Data

  • Poor data quality
  • Automated data capture and alternative/additional data sources limited or absent
  • Problems identifying which data is predictive and when
  • Limited automation of customer data validation and verification
  • “Stove piped” data in diverse systems, difficult to access
  • Issues with integrating to 3rd party data and multiple data sources
  • Management of knowledge repositories not consistent or absent
  • Difficult to get data when needed

Decisioning

  • Long time to get changes to market
  • Time consuming to change/implement decisioning logic/policies/rules/models/workflows as part of customer and product lifecycle management
  • Problems with integrating/adapting decisioning logic with user interface, loan management system (LMS) and other modules
  • Sub-optimal acceptance, conversion, retention, default, and fraud rates
  • Restrictions in creating and testing new business workflows, rules, policies, models, data verification logic, long development cycles
  • Lack of transparency/traceability of decisions

Operations & Business Process Management

  • Lack of flexibility to modify product types, adapt customer & product lifecycle processes
  • Modifying acquisition channels slow, (e.g. Web, mobile, USSD, voice/IVR applications, SMS, chat, etc.)
  • Enhancing system functionality and changing / automating business process workflows difficult and time consuming
  • Too much manual effort in the customer & product lifecycle
  • Limitations in automating LMS processes were often sited, e.g. “possible to select a wrong option” or errors due to numerous data field entries
  • Limitations for online/mobile lending systems where customer volumes are significant/growing
  • Performance, scalability and fault tolerance issues
  • Lack of STP (straight through processing) support. A number of LMS’s and decision engines could not be utilized for STP online/mobile lending systems
  • In-house platform struggles to get timely and necessary changes delivered. On-line business, needs rapid reaction, and “agile” iteration, and so is constrained
  • Maintenance/development costs too high
Solution

Embedded in Acquired Insights adaptive fintech platform is deep fintech knowledge, behavioral analytics, artificial intelligence and robotics. The platform adapts to customer behavior dynamically and intelligently so that you can fully automate your digital product.

Acquired Insights adaptive fintech platform is a SMART system, one of the most powerful financial platforms in global use today. It has been designed to help transform products from off-line to digital. Its AI/ML systems and workflows intelligently manage the entire customer, portfolio and product lifecycle.

Acquired Insights operates one of the largest Artificial Intelligence / Machine Learning research labs focused on fintech automation:

  • personalizing products and services
  • optimizing operational processes
  • improving customer experience
  • ongoing predictive models tuning
  • customer verification
  • dynamic customer journey
  • portfolio performance

Workflows that are "Intelligent" Controlling Process & Decisioning

Acquired Insights adaptive fintech platform intelligent workflows allow companies to market and sell digitally, much as you would face to face. Our platform:

  • adapts to individuals
  • dynamically adapts your sales approach in real time, as you learn more about your customer
  • tests new approaches, models, techniques, strategies, products, offers etc.
  • experiments, implements and adapts those that produce the best results
  • monitors and compares performance and further tune
  • gracefully handles a large range of predictive data and Big Data volumes
  • customizes and optimizes your business processes
  • manages and executes real-time decisioning

Acquired Insights adaptive fintech platform captures and supplies a large range of predictive data. This is important, as it has a major effect on:

  • personalizing products and services
  • optimizing operational processes
  • improving customer experience
  • ongoing predictive models tuning
  • customer verification
  • dynamic customer journey
  • portfolio performance

Credit Scores & Bureau Data aren't Enough

Most of our clients’ systems did not capture all the data required. Acquired Insights adaptive fintech platform does.

Typically banks/finance organizations use credit scores and traditional credit bureau data where available. This data is useful, but far more is needed, from a wider range of sources, (e.g. behavioral, application, historic, device, social, MNO, unstructured, bank statements, lead generator, marketing, 3rd party and many others). This is necessary to obtain a full “picture” of the client in order to optimize each step of the sales process and relationship.

Acquired Insights adaptive fintech platform was designed from the outset to collect these. More importantly, it knows which features to collect, which are predictive, for what and when. This is crucial and represents over 300 man years of R&D.

With Acquired Insights adaptive fintech platform organizations can manage the entire customer relationship from the moment a lead is referred to their site. When we say “manage”, we also mean “intelligently automate,” with minimal human intervention.

It’s not just about credit scores, however. Predicting default/risk is, of course, important. To fully automate a digital product and scale requires calibration of a larges range of collaborative models, looking at each aspect of the client relationship and behavior. In certain areas of the product lifecycle, model ensemble and boosting techniques are required to optimize profitability.

Going Global

Acquired Insights adaptive fintech platform covers our global experience. So you can rapidly expand your use of the platform across states, countries and continents, even where little or no credit agency/bureau data exists. (NB Acquired Insights has specialized in this throughout the Americas, Europe, Africa and Asia).

Overall Improvements:
  • Decrease the Number of Defaults by 28%
  • Straight Through Processing by 50%+
  • Reduce Fraud Rates from 2% to 0.1%
  • Increase Acceptance by 33%
  • Increase Renewal Rate by 22%
  • Increase Profitability by 28%+

© Zoral Limited 2017 All rights reserved  

© Copyright www.aiinc.cloud