top of page

AI Case Study

An anonymous global insurance company analysed customer contact recordings and claims data to drive sales yields using natural language processing

Insurance companies have a wealth of unstructured data, such as sales calls, that could provide valuable insights if understood, categories and actioned. A global insurance company worked with Re:Infer to analyse contact recordings, claims data and employee feedback to identify key issues and sentiment using natural language processing. Focusing on sales intelligence they were able to increase sales yields by identifying behaviours that lead to successful up-sell and cross selling. They also improved product intelligence, claims and employee insights.


Financial Services


Project Overview

Re:Infer, a UK natural language processing solution provider, worked with a global insurance business. They provide "the capability to listen and understand communications data at scale. Re:infer can ingest data from multiple sources and provides real time annotations of comms data to trigger and inform downstream processes (e.g. RPA processes and Marketing Automation services triggered via our API)."

Reported Results

They claim improvement in "sales intelligence and optimization that included

* Increased sales yields by identifying behaviours that lead to successful up-sell and cross-sell activity
* Identify and reduce barriers to sales efficiency through statistical analysis of sales cadence
• Pricing intelligence – calls analytics identified phrasing and promotion of price offered key determinant to close of sale"

They also claimed advances into product, employee and claims insights.


Advanced natural language processing capabilities of Re:infer team.





Insurance companies are awash in unstructured data. Analysing calls, claims and employee feedback could provide valuable insights into sales effectiveness, product market fit and even employee concerns.



"Call recordings - 100,000 broker to underwriter calls. 12,000 claims on property business line. 100,000 unique employee feedback verbatim. 12,000 unique email complaints."

bottom of page