Dezerv Enhances Wealth Management with Sarvam's Post-Call Analytics
By leveraging Sarvam's Speech to Text technology, Dezerv transforms client conversations into searchable data, enhancing insights and fostering trust in wealth management. This strategic move meets the growing demand for personalized financial advice in India's expanding market.
Key Facts
- Dezerv manages ₹17,000cr+, highlighting rapid growth in India's wealth management sector.
- Sarvam's 20% reduction in Word Error Rate enhances client trust through improved communication.
- In-house transcription allows real-time insights, a competitive edge over traditional firms.
- Processing 3.6 lakh minutes monthly indicates high operational efficiency and scalability potential.
- Customization of services through data-driven insights positions Dezerv as a market leader.
Summary
Summary
Dezerv, a wealth management firm in India, faced challenges in extracting actionable insights from client conversations. By implementing Sarvam's Speech to Text technology, Dezerv transformed thousands of calls into searchable transcripts, significantly enhancing their understanding of client needs. As a result, they improved transcription quality by approximately 20% and processed nearly 3.6 lakh minutes of speech in one month.
Background
Founded in 2021, Dezerv operates in the wealth management industry, focusing on creating customized investment portfolios for clients with a minimum investment of ₹50 lakhs. By June 2026, the firm managed over ₹17,000 crore across various investment services and employed a team of over 650 across five cities. The relationship manager (RM) plays a crucial role in client interactions, conducting conversations in multiple languages over video conferencing.
Challenge
Dezerv recognized that the conversations between RMs and clients were a rich source of insights, capturing client concerns and preferences. However, the lack of a reliable system to convert these conversations into searchable data meant that much of this valuable information remained untapped. The previous outsourced transcription process was inadequate, particularly for Indic languages, and did not allow for quality audits or comprehensive data queries.
Solution
Dezerv integrated Sarvam's Speech to Text technology through its SDK, enabling three primary use cases: transcription of video meetings, transcription of phone calls, and live note-taking for RMs during client conversations. With client consent, recordings were processed using Sarvam's model, which included speaker diarization to create structured transcripts. This system ensured that no client-identifying information was stored, adhering to Dezerv's data governance framework.
Results
With the new pipeline, Dezerv converted client conversations into searchable transcripts within hours. In one month, they processed approximately 3.6 lakh minutes of speech, including around 3.2 lakh minutes of video meetings and 40,000 minutes of phone calls. The transition to Sarvam's technology led to a 20% reduction in Word Error Rate compared to the previous setup, providing Dezerv with better visibility into performance and quality.
Key Insights
The deployment of Sarvam's technology illustrates the importance of integrating advanced speech recognition systems in client-facing industries. By leveraging accurate transcription and speaker identification, firms can enhance their understanding of client needs and improve service delivery. Additionally, maintaining data governance is crucial when handling sensitive client information.
Customer Testimonial
“A Relationship Manager conversation is where trust is built or lost. For years the most important thing in our business was also the thing we could see the least of. Sarvam turned thousands of those conversations into something we can learn from.”
— Sandeep Jethwani, Co-founder, Dezerv
Entities Mentioned
Companies
Products
Technologies
People
Organizations
Key Concepts
Definitions
- Speech to Text
- A technology that converts spoken language into written text, often used for transcribing conversations.
- Diarization
- The process of distinguishing and labeling different speakers in an audio recording.
- Word Error Rate (WER)
- A metric used to evaluate the accuracy of speech recognition systems by measuring the percentage of words incorrectly transcribed.
- Character Error Rate (CER)
- A metric similar to WER, but it measures the accuracy of transcription at the character level.
- Zero-Retention AI Processing
- A data processing approach that ensures no client-identifying information is stored after processing.
Use Cases
- →Transcription of video meetings
- →Transcription of phone calls
- →Live note-taking for Relationship Managers during client conversations
Frequently Asked Questions
What is the main benefit of using Sarvam for transcription?
Sarvam provides high accuracy in transcribing Indian speech, including code-mixed conversations, which is crucial for wealth management discussions. This allows Dezerv to gain valuable insights from client interactions.
How does Dezerv ensure client data privacy?
Dezerv manages client data under a strict data governance framework and employs Zero-Retention AI Processing, ensuring that no client-identifying information is stored with Sarvam.
What improvements has Dezerv seen since integrating Sarvam?
Dezerv has significantly improved transcription quality, reducing the Word Error Rate by around 20% compared to their previous setup. This has enhanced their ability to analyze client conversations effectively.
How does the transcription process work?
Dezerv processes client call recordings through Sarvam's Speech to Text model, which generates structured transcripts with speaker diarization. This allows Relationship Managers to easily access and review past conversations.
What future developments does Dezerv anticipate?
Dezerv aims to continue enhancing its wealth management services by leveraging insights from client conversations. As conversation volumes grow and Sarvam improves its models, the value of the insights will compound.