In the fast-paced world of wealth management, where every conversation can be a goldmine of insights, Dezerv is revolutionizing the way firms like theirs operate. Founded in 2021, Dezerv aims to bridge the gap between India's growing wealth and the expertise needed to manage it. With a focus on customized investment portfolios and a team of 650+ across five cities, Dezerv is not just a wealth management firm; it's a beacon of innovation in the industry. But what truly sets Dezerv apart is its approach to harnessing the power of its relationship managers (RMs) and the insights they provide. And that's where the story of Sarvam comes in.
Dezerv's relationship managers are the heart of the firm, and their conversations with clients are a treasure trove of information. However, these insights were locked away in audio recordings, inaccessible and difficult to analyze. The previous transcription pipeline was outsourced, which meant Dezerv couldn't audit its own quality, and the transcripts were often inadequate, especially for Indic languages. The transcription process itself was a challenge, as Dezerv's clients and RMs speak a mix of languages and use technical terms, making accuracy critical, especially for numbers.
This is where Sarvam comes in. Sarvam's Speech to Text platform was chosen after a rigorous benchmarking process, and it consistently delivered the strongest performance. What makes Sarvam stand out is its accuracy on real Indian speech, with models trained on over a million hours of Indian audio, including code-mixed conversations, regional accents, and telephony recordings. This is exactly the kind of conditions Dezerv's calls present, and it's a game-changer.
But Sarvam's capabilities don't stop there. Its built-in speaker diarization automatically separates different speakers throughout a conversation, which is crucial in wealth management to know who said what. And Sarvam's ability to scale without operational complexity means Dezerv can submit large volumes of conversations asynchronously and receive structured, speaker-separated transcripts back, without the need to build or operate its own transcription systems.
The integration of Sarvam through Dezerv's SDK has transformed the way the firm operates. With client consent, recordings are processed via Sarvam's Speech to Text model, using the Batch API with speaker diarization. Every conversation is converted into a structured transcript that captures who spoke, what they said, and when they said it. This searchable record of the client relationship gives RMs the context they need to carry each client relationship forward, and it's a game-changer.
The impact of this pipeline is significant. Dezerv now turns every client conversation into a searchable, speaker-separated transcript within hours. Last month alone, the company processed close to 3.6 lakh minutes of speech through Sarvam, including around 3.2 lakh minutes of video meetings and 40,000 minutes of phone calls. At this scale, Dezerv has also significantly improved transcription quality, reducing Word Error Rate by around 20% compared with its previous setup.
Looking ahead, Dezerv is building the wealth management firm India's professionals were never offered: expert-led, technology-first, and accountable for outcomes. Learning from every client conversation is central to this vision, and as conversation volumes continue to grow and Sarvam advances its speech models for Indian languages, the value of that institutional knowledge will only compound. In my opinion, Dezerv's partnership with Sarvam is a testament to the power of innovation and the potential for technology to transform the wealth management industry. It's a fascinating development that will shape the future of the industry, and I can't wait to see what comes next.
