A large banking client is required by law to provide tax relief at source (TRS) in the form of a refund or reduction in mortgage payment to its customers. They were dealing with a large number of exceptions that required manual intervention resulting in significant loss of productivity and cost.

Infosys Nia’s machine learning model was trained with historical exception and tax account data. A set of 10 business rules were used to classify accounts and correlation analysis was performed to root cause exceptions. Exceptions were highlighted in a unified dashboard view. The prediction model predicted the likelihood of an exception on each account, and automatically resolved exceptions when they occurred.

Benefits:

• Cost Reduction of € 250K per year
• Improved Customer Satisfaction
• Overview of Exceptions with Drilldowns
• Prediction and Automated Resolution of Exceptions

Payment disputes are an important reason for increased DSO, which when exacerbated by an incorrect collection strategy, not only increases DSO, but also results in poor customer experience.

Infosys Nia brought together data from all the different systems – ERP, Order Management, invoicing – within the organization and stitched it with external macro-economic and behavioral data to create a complete customer risk profile. This was then aggregated to predict the total account value at risk.

Benefits:

• Customized collection strategy
• Expedited resolution of disputes through automated root-cause
• Improved cash flow forecasting
• Identify anomalies to predict and prevent disputes

One of our clients operates with multiple entities under multiple contracts that govern how business is conducted. They were dealing with a extended lead times on exception handling due to the manual process of reviewing multiple contracts and understanding contract rules to resolve the exception.

Infosys Nia ingested all contract agreements to create a knowledge model that captured rules, exceptions and resolutions. The model responded to incoming queries and resolved exceptions based on information captured from the labor contracts. A conversational interface provided natural language query into the knowledge models.

Benefits:

• Conversational interface to query and interact with knowledge model
• Dynamic knowledge model that learns and evolves
• Improved staff productivity with on-demand access to operational knowledge

The order-to-activation cycle is a critical business process in the Telecom industry. Order volumes are large, orders can be complex with bundled services and several applications tracing thousands of unique paths. Often legacy systems are adapted as-is to newer service categories. This leads to multiple failure scenarios resulting in higher costs, increase in cycle time and delay in activation.

Benefits:

• Decrease in cost per order by as much as 25%
• Improvement in on-time delivery by up to 40%
• Superior user experience across the ‘Order-to-Activation’ process