Product Owners
Insurance
Industry
Northbrook, USA
Location
60,000+
Company size
What interested me in super.AI is not only labeling for a specific project in a small siloed environment but how that knowledge can be leveraged at an Enterprise level and how that can help to scale the work with the support of AI and that's what got me very excited.
Senior Manager

Starting point: Using AI for improving customer satisfaction

Our client is a top American Insurance company with decades of operations and expertise in car insurance. The company contacted super.AI to help with a scalable processing of their customer calls, which they wanted to use for an active learning algorithm. The ultimate objective was to increase customer satisfaction and retention by matching customer sales queries with the most relevant information.  They had been searching for a valuable partner which had multi-modal labeling and active learning capabilities, so that they could assist with ML models analysing customer sales calls. They were drawn to the super.AI solution due to its end to end capabilities, from labelling to model deployment.

Scaling sales calls processing with super.AI

The customer had very stringent privacy requests and wanted to limit the access to the customer calls to their internal labelers. They were able to do that thanks to the super.AI option to add own labelers to the project.

The customer labeling team started by employing our audio transcription data program to transcribe all the sales calls audio files into text. The second step was to use our query intent matching data program. They inputted the search queries and possible intents to each query and produced an outcome file which had a score matching of intent for each query, on a scale of 0.1 (no connection to 1 (perfect match). In parallel, the customer's labeled data was automatically inputted into an active learning algorithm.

Results

By using the super.AI NLP solution, the company was able to achieve several benefits:

  • maintain strict data privacy requirements for processing of sales calls
  • use internal team of labelers
  • process a large amount of sales call quickly
  • in parallel, input the labeled data into an active learning algorithm.

By running the data labeling and the training of the active learning algorithm in parallel, the company was able to significantly reduce timelines for the deployment of the ML model, while ensuring a high degree of quality within their cost parameters. Ultimately, the customer was able to significantly increase the relevance of recommendations provided for main customer queries during the sales calls, thereby increasing customer satisfaction, retention and revenue.

Related Case Studies