Listen to the article
The Bengaluru startup is shifting beyond cheaper automated calls to embed AI in insurance and finance processes, where compliance, reliability and measurable outcomes may offer a stronger business moat.
On the Prime Venture Partners Podcast, Gaurav Ranjan spoke with Desible.ai co-founders Uttam Tiwari and Omkar Raikar about a proposition that is becoming more compelling as voice AI gets cheaper: the real value, they argued, lies not in sounding human, but in embedding AI inside the workflows that keep regulated businesses moving. Prime Venture Partners’ episode page says the discussion focused on building agentic AI for banking, financial services and insurance, with attention to compliance, security, production deployment and the realities of enterprise sales in India.
The founders’ thinking was shaped by an earlier logistics brokerage project in which AI handled calls to truck drivers for availability and verification. According to the podcast discussion, the client was able to shrink a contact centre that had employed roughly 20 to 25 people down to just three, a result that convinced the pair that voice technology could change how enterprises operate. Yet after running proof-of-concept projects with a dozen or so businesses in India, they concluded that conversation alone was not enough to build a defensible company. As Uttam Tiwari put it on the podcast, “Voice is a commodity. I do agree, but at the same time, workflows are not because workflow goes and sit deep inside organizational processes and that’s where you get the moat.”
That insight pushed Desible towards insurance, where a customer interaction often feeds directly into renewals, persistency, claims, sales and servicing. YourStory reported that the company is now positioning itself as an agentic workflow orchestration platform for regulated enterprises, with insurance as its first major focus. The move fits a broader market shift described by the company and its backers: not just automating calls, but inserting AI into repeatable business processes where performance can be measured in outcomes rather than novelty.
The gap between an impressive demo and a system that can run inside a regulated organisation emerged as one of the central themes of the conversation. Uttam Tiwari said conversion from proofs of concept to live deployments among the companies he has seen is below 35% to 40%, while Omkar Raikar described the extensive scrutiny that Indian insurers and banks impose before allowing any new software into production. That scrutiny can include information security questionnaires running to hundreds of questions, covering issues such as data residency, infrastructure and auditability. Desible, according to the founders, spent much of its early funding on compliance infrastructure rather than voice features alone.
The technical demands are not limited to paperwork. Omkar Raikar said insurance conversations require a level of consistency that leaves little room for improvisation, and gave examples of an AI agent handling interruptions, multilingual exchanges and medical questionnaires while still staying within policy and compliance boundaries. A system may need to recognise that a customer has interrupted a question, identify a medicine mentioned mid-call, verify it against a database and then continue with the right follow-up. In that sense, the product challenge is not merely to generate speech, but to maintain context and operate reliably across the full exchange.
The economics of the sector are changing almost as quickly as the technology. Gaurav Ranjan noted during the discussion that human contact centres can cost roughly ₹20 to ₹30 per minute, while early voice AI systems were priced at around ₹5 to ₹7 a minute. Today, competitive offerings have fallen to roughly ₹1.5 to ₹3 per minute. Desible says it has reduced its own underlying call-execution costs from around ₹7 to ₹8 to below ₹1.5 to ₹2 through orchestration. But that also creates a pricing dilemma for the industry: if the core infrastructure keeps getting cheaper, providers must find another way to capture value. The company’s answer is to charge for workflow execution rather than treating each call as a standalone transaction.
The shift in buyer behaviour is just as important. A year ago, enterprise customers were largely judging voice AI on sound quality, latency and how naturally a system could switch languages. Now, Uttam Tiwari said, they are comparing AI with human workers on contact rates, conversion, lead qualification and the accuracy of data captured after a call. That changes the sales process as well. “Your champion may not be your buyer,” he warned, pointing to the reality that enthusiasm from one business stakeholder is rarely enough to secure a deployment when technology, compliance, procurement and security teams all have a say. Omkar Raikar suggested that in India, navigating those internal dynamics can matter even more than the product itself.
That commercial reality helps explain Desible’s account-based expansion strategy. The company begins with one workflow, such as sales funnel management, and then extends into welcome calls, persistency and inbound servicing once trust is established. According to the founders, it has already identified more than 20 workflows within renewals alone. Inc42 reported that the Bengaluru-based startup recently raised ₹32 crore in a round led by Prime Venture Partners, with participation from existing investor Invention Engine, and said the capital would be used to expand its agentic AI capabilities, improve compliance infrastructure and develop more products for financial services. Entrackr independently reported the same funding round.
The wider ambition goes beyond voice altogether. Desible’s company materials describe a conversational AI platform for banking, finance and insurance, and the founders said on the podcast that they want to build an operating layer that can support insurers, brokers and third-party administrators across multiple workflows. For them, the prize is not a more convincing synthetic voice, but infrastructure that helps large organisations make better decisions, improve customer outcomes and run more efficiently. As Uttam Tiwari summed it up, the industry’s current state is one in which “POC is overhyped but production is underhyped.”
Source Reference Map
Inspired by headline at: [1]
Sources by paragraph:
- Paragraph 1: [2], [3]
- Paragraph 2: [1], [3]
- Paragraph 3: [3], [4], [5], [6]
- Paragraph 4: [1], [2], [3], [6]
- Paragraph 5: [1], [3]
- Paragraph 6: [1], [2], [3], [4], [5], [7]
- Paragraph 7: [1], [4], [5], [6], [7]
Source: Fuse Wire Services
Access Our Exclusive Research
In-depth analysis and actionable data to keep you ahead of the curve.
Never Miss an Update. Get our latest research delivered straight to your inbox.