Ringg raises $10M from Peak XV to push voice agents beyond phone calls

Indian voice-AI startup Ringg has closed a $10 million extension (about 37 million shekels) to its Series A, led by Peak XV Partners. The new money brings the total round to $15.5 million on top of the $5.5 million raised earlier. The company, which launched as a text-to-speech venture called DesiVocal, abandoned training its own speech models when the compute bill became unsustainable and rebuilt as an application layer that deploys voice agents for enterprise customers. Its first design partner was the fintech Cred; Flipkart, Practo, Groww and PolicyBazaar have since signed on. Ringg says its platform now fields 20 million call attempts a month.

Co-founder Siddharth Tripathi says the early focus was high-volume, low-complexity work — outbound dialing, lead qualification, loan collections — but those use cases turned out not to be sticky and collapsed into pure price competition. The company is now targeting richer workflows: scheduling clinic appointments, recovering abandoned e-commerce carts, running KYC checks inside fintech apps. Tripathi notes the voice agent is already live across 1,200 clinics on the Practo health app, helping patients book visits or follow post-visit instructions.

Voice still accounts for more than 70% of activity, but Ringg has begun rolling out chat, WhatsApp and browser-based support channels — Shell is one early example. "We're trying to position ourselves as a platform for agents that drive outcomes or complete tasks, not just as enterprise voice agents," Tripathi said. The U.S. go-to-market strategy bypasses direct sales in favor of partnerships with Global Capability Centers — the offshore hubs multinational corporations increasingly rely on for back-office and support work — so Ringg can sell automation alongside human teams.

Tripathi adds that Ringg is building its own speech-recognition and generation models with the long-term goal of owning the full voice stack, including infrastructure and deployment. For now, though, the cost is prohibitive and the product operates as an orchestration layer that routes tasks to different models depending on the use case. Peak XV partner Rishen Kapoor said Ringg's origins as a research lab that built models from scratch shows up in its ability to handle complex enterprise workflows end-to-end — merchant onboarding, L1 and L2 support — with quality and consistency.

The competitive field in India is crowded. Model makers such as Deepgram, ElevenLabs, Cartesia, Sarvam and Smallest.ai are vying for the base layer; orchestration-focused startups including Bolna and Blue Machines are chasing the middle layer; and vertical players like Gnani and Arrowhead specialize in finance. Those three strata — model builders, orchestrators, and application-layer players all trying to lock in enterprise workflows — are the real story. The money and the defensibility increasingly sit with whoever owns the customer relationship.