Improving Customer Service with intelligent Automation for the largest South African Long Term insurer

Insurance | Customer Servicing | Intelligent Automation

Case Study Details

Whilst much of the focus around Artificial Intelligence has been around new ways of interacting with customers, there is a huge opportunity in applying intelligent automation to traditional customer service channels such as email which still makes up the bulk of messaging for most large organisations.

Challenge

The largest long term insurer in South Africa required a more efficient manner in which to handle their mountain of incoming email messaging in there main contact centre.

challenge

Solution

To address this, we embarked on a project to determine the feasibility of implementing a custom natural language understanding machine learning model and associated pipeline to read each email, extract the relevant data attributes and then action each email in the respective downstream application.

Approach

  • Classification Model Training: We trained a natural language classification model to identify the nature (type of service request) of each incoming email.
  • Natural Language Extraction: We trained a natural language model to extract all relevant data attributes from each mail e.g. name, contact details, etc.
  • Pipeline Engineering: We developed a bespoke email ingestion and understanding pipeline which integrated to the organisations mail server and downstream applications.
approach

Results

The implementation of our intelligent email automation solution resulted in:

  • Improved Operational Efficiency: Customer service requests were dealt with in effectively real time as opposed to hours and days.
  • Improved Customer Satisfaction: Customers were receiving email responses in effectively real time as opposed to hours and days which greatly increased their CSAT scores.

Lessons Learnt

  • Integration Complexity: Large organisations have brownfields operational systems which are numerous and complicated to integrate to.
  • Organisational Resistance: Legacy pockets of technology and their associated representatives can be very resistant to changes especially those that compromise their existing role and function in the organisation.
  • Executive Stakeholder Alignment: Getting executive buy in from the start would be the number one lesson learn in order to smooth the implementation process.
  • Return on Investment: Traditional AI initiatives are not purely cost saving exercises and depending on this to justify ROI is challenging. One has to look to the business value derived from increased operational processing times and the like.
lessons-learnt

Conclusion

This project proved beyond any doubt that the latest natural language understanding AI technologies can be leveraged to great effect to transform the level of customer service achievable in the modern contact centre.