For a leading car dealership, we are analysing 1,000+ calls daily for Customer sentiment and feedback, to drive sales
What we were solving for
Customer feedback: Manually analysing large volumes of audio can be challenging, whilst only analysing a small sample can often result in critical insights being lost
Customer sentiment: By understanding the level of interest, our Client was seeking guidance on which lead to offer a white-glove experience to, to expedite the sales cycle
OptiSol solution
We converted audio to text using AWS Transcribe and segmenting the text spoken by different participants
A natural language processing pipeline was built to assess the tone of the call and score it
Feature selection was applied to only select the most important “features” from a call, to analyse further
We built a classification model to predict whether a lead is likely to convert into a sale