Solution Overview

Reducing past-due invoices and improving working capital with predictive collections

  • Facebook
  • Twitter
  • Linkedin
  • Email

Volatility in customer payment behavior and changes in local and global business environments can have a significant impact on a company’s ability to manage business-to-business (B2B) credit and risk in its accounts receivable (AR) function. Current approaches can fail to address these issues proactively, leading to higher levels of days sales outstanding (DSO) and tying up working capital.

B2B collections processes traditionally follow AR aging-based approaches that are designed for stable economic times. But with volatility in political and local market conditions, collections strategies and portfolio management must become less reactive by adopting probabilistic algorithms that predict past-due invoices and AR risk, and enhance segmentation.

Take a copy for yourself

Download PDF

The challenge

For many global organizations, B2B collections strategies are based on the aging of AR—these organizations have limited segmentation capabilities, sporadic access to external data, and poor portfolio management. Existing ERP tools and collection workbenches struggle to store and analyze large data-sets with multiple variables and complex data structures. Organizations also often take a static approach to collections strategies due to legacy processes and relationship maps.

These challenges impact companies in a number of ways:

  • Productivity:
    • Our experience indicates that companies waste between 10% and 40% of their collections spend due to suboptimal portfolio definition and segmentation
    • Account coverage per AR collector is approximately 65% to 70%, which leaves up to 30% of AR untouched each month
  • Days sales outstanding: High DSO levels can range from 35 to 50 days in different regions
  • Working capital: For large, multinational organizations with billions of dollars in revenues, each day of outstanding AR can result in millions of dollars in lost opportunity cost
  • Profitability: Up to 5% of earnings before interest and taxes is locked up in collectible invoices, past due

Scientific and statistical approaches for predicting customers’ payment behavior are often missing in B2B collections.

As this payment knowledge is typically held by individual collectors, and is based on their experiences and expertise, they make decisions according to rules of thumb or intuitive judgments, which result in: 

  1. Low collections effectiveness for managing past-due accounts and invoices
  2. High dependence on collectors’ experience of customer payment behavior, which increases the knowledge gap when AR professionals move to new roles
  3. Flat performance on past-due aging as there is limited learning from historical data

Without a dynamic feedback loop between AR collections risk and credit risk management, organizations face challenges where credit review and assessment are irregular, and depend on aging threshold-based triggers and limited external credit appraisal data. The consequences are:

  1. 1% to 3% of annual bad-debt write-offs as credit-risk exposure is not controlled by a dynamic assessment of a customer’s credit performance
  2. Revenue leakage of 0.5% to 1% due to a lack of timely intervention when assessing credit improvement and deterioration

Genpact solution

Genpact’s cloud-based Dynamic Credit and Collections Engine leverages advanced digital technologies, such as machine learning, natural language processing, and big-data analytics, to drive predictive modeling and accurately forecast past-due invoices and AR risk. Machine learning enables a continuous cognitive assessment of collections portfolios and strategy effectiveness by identifying patterns across vast data sets. It helps predict customer payment behavior, and manage customer credit risk in real time.

The engine’s capabilities include:

  • Integration of multiple data forms and sources to deliver a statistical view of customers’ credit health, including:
    • Historical structured payment data, such as payment history and frequency of late payment/disputes, which are extracted from ERPs, collections workflow, reporting solutions, etc.
    • External structured customer data, such as liquidity ratios, financial stress scores, and credit scores to determine customers’ profitability and business health
    • External unstructured data, such as legal and political news, customer reviews and feedback, political disturbances or other events that could impact customers’ businesses and deals won
  • Statistical predictive analytics to both anticipate customer payment behavior; predict potential late payment of an invoice and its probable aging. This enables dynamic, multidimensional customer segmentation
  • A dynamic risk-based segmentation model that builds on a credit and AR-risk scoring framework to define multi-dimensional customer segments and deliver an efficient strategy. The model refreshes as new information becomes available (see figure 1)
  • Dashboard of customer and invoice past-due risk and scoring status. The dashboard guides credit and collections teams with segmented actions to drive customer credit and collections response workflows
  • A decision-support collections workflow based on customer size, AR at risk, channel, and other parameters to enable more effective customer segmentation, strategy definition, and portfolio management

Figure 1: Dynamic risk segmentation model



Genpact’s Dynamic Credit and Collections Engine is powered by our Lean DigitalSMapproach that harnesses design-thinking methods, Lean practices, and advanced digital technologies. The solution’s sophisticated analytical models and process workflows have been shaped by our knowledge of running collections operations for more than 90 clients.

Organizations experience benefits along two tracks:

  1. Probabilistic prediction of invoice past due and AR risk prior to invoice due date to anticipate:
    • Which of the open invoices will be paid late?
    • The most probable aging of overdue invoices?

    The answers to these questions allow collections teams to adopt proactive strategies that reduce aging and late payment.

  2. Statistical and scientific portfolio segmentation and strategy optimization that maximizes the ROI on collections-spend by integrating collectors’ experiences and domain expertise with advanced analytics. This enables informed decision-making to determine:
    • Where organizations should focus collections efforts
    • Which team member performs the collection
    • How to maximize collections while minimizing the cost to collect

The Dynamic Credit and Collections Engine can be easily configured to work alongside existing systems. It integrates digital technologies and predictive analytics to anticipate customer payment behavior and assess real-time credit risk, which supports segmented and customer-focused collections strategies.

Key benefits include:

  • Improved DSO by 5% to 15%
  • Reduced past-due invoices to less than 1% and sustained performance
  • Strengthened working capital management due to improvements in DSO and past-due invoices
  • Increased collections efficiency by up to 40%
  • Reduced collections costs by 20% to 30% by adopting

Genpact’s credit and collections engine, organizations are meeting their core credit and collections objectives.

Visit our Finance and Accounting solutions

Learn More