About the Industry

Telesales services involving sales of products or services over a phone call is a vast market with a global value of approximately US$ 27 billion in 2022 and is expected to grow to US$ 55 Bn by 2029. Call centre solutions originating from India are a widespread phenomenon and most of these services are for the vast consumer market in India with calls happening in Indic languages like Hindi, English, Bengali, Tamil, etc. Hence countless call recordings are generated with millions of conversations happening in a day across various industries like Financial Services, Healthcare, Insurance, Hospitality, Marketing etc.

Why Speech Recognition is Important in Telesales?

Despite the substantial volume of call recordings generated on a daily basis, the telesales industry remains relatively unexplored in terms of engineering innovations and AI applications. Since Speech Recognition is generally a pre-requisite for most of these applications and its quality in Indic languages often serves as a limiting factor.

These call recordings can enable organisations to innovate using Artificial Intelligence and build applications like lead prioritisation, mining product feedback, automated training of telesales agents, language or grammar assessment, conversation analytics and many more, to provide best-in-class services to their customers and visibility to their business.

Problem Statement

At SquadStack, we provide fully managed telecalling services to our customers across various industries like Financial Services, Healthcare, Insurance, E-commerce, etc. In the process, we generate thousands of hrs of call recordings in a day while calling leads all across India. We realised a need for customised Automatic Speech Recognition (ASR) build specifically for the telesales industry in the Indian context to empower many of the AI applications in this industry.

Outline for the talk

• Why behind the problem?
• What are we solving? The proposed workflow in the telesales industry.
• Speech Recognition System Architecture
• Implementation and deployment of service at scale
• The impact generated in terms of cost, quality and unlocking downstream applications
• Benchmarking across available services
• Future work and conclusion