Call center ai · Production

Verisk scales customer sales call processing with super.AI NLP and active learning

The problem

Verisk needed a scalable, privacy-compliant solution to process large volumes of customer sales call recordings and feed the results into an active learning algorithm, with the goal of improving recommendation relevance and increasing customer satisfaction and retention.

Workflow diagram · grounded in source
1
Audio transcription of sales calls
ai_action
“employing our audio transcription data program to transcribe all the sales calls audio files into text”
2
Query intent matching and scoring
ai_action
“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 (perf…”
3
Parallel active learning input
feedback_loop
“in parallel, the customer's processed data was automatically inputted into an active learning algorithm”
4
Improved sales recommendations
output
“significantly increase the relevance of recommendations provided for main customer queries during the sales calls”
Reported outcome

By running data processing and active learning training in parallel, Verisk significantly reduced ML model deployment timelines while maintaining quality within cost parameters, and significantly increased the relevance of recommendations for customer queries during sales calls, driving higher customer satisfaction, retention, and revenue.

Reported metrics
ML model deployment timelinesignificantly reduce timelines for the deployment of the ML model
Recommendation relevancesignificantly increase the relevance of recommendations
Customer satisfaction, retention, and revenueincreasing customer satisfaction, retention and revenue
Reported stack
super.AI NLP solutionaudio transcription data programquery intent matching data programactive learning algorithm
Source
https://super.ai/case-studies/large-us-insurer
Read source ↗

Frequently asked questions

What did this team achieve with this AI workflow?

By running data processing and active learning training in parallel, Verisk significantly reduced ML model deployment timelines while maintaining quality within cost parameters, and significantly increased the relevan…

What tools did this team use?

super.AI NLP solution, audio transcription data program, query intent matching data program, active learning algorithm.

What results were reported?

ML model deployment timeline: significantly reduce timelines for the deployment of the ML model; Recommendation relevance: significantly increase the relevance of recommendations; Customer satisfaction, retention, and revenue: increasing customer satisfaction, retention and revenue (source-reported, not independently verified).

How is this call center ai AI workflow structured?

Audio transcription of sales calls → Query intent matching and scoring → Parallel active learning input → Improved sales recommendations.

Related call center ai cases