Startups

Indian startup Ringg raises $10 million to expand voice AI agents

Ringg secured $10 million from Peak XV Partners as an extension of its Series A round, bringing the round’s total to $15.5 million. The company is shifting from simple automated calls to executing more complex enterprise tasks through voice, WhatsApp, chat and browsers.

2026-08-25
4 min read
1,215 views
certi.news Editorial Team
Indian startup Ringg raises $10 million to expand voice AI agents

Indian startup Ringg, which specializes in voice AI agents, has raised $10 million from Peak XV Partners as an extension of its Series A round, which raised $5.5 million earlier this year. This brings the round’s total funding to $15.5 million, as the company bets on consumers in India continuing to rely on phone calls to communicate with businesses.

The funding round comes as Ringg says it handles 20 million calling attempts each month. According to a recent Truecaller study cited in the article, more than 76% of consumers in India prefer speaking with businesses by phone, creating an opportunity to automate support and outreach calls using voice AI.

From mass calls to task execution

Ringg began as DesiVocal, a company developing text-to-speech technologies, but changed direction after finding that training its own voice models was expensive. Rather than focusing on the model layer, it shifted to building voice agents for enterprises. Financial company Cred was its first customer, followed by Indian companies such as Flipkart, Practo, Groww and Policybazaar.

The company is now focusing on workflows more complex than outbound calling, lead qualification and loan collection. New use cases include booking appointments at clinics, recovering abandoned shopping carts in e-commerce, and customer onboarding and identity verification procedures, known as KYC, in fintech applications.

Co-founder Siddharth Tripathi said Ringg’s voice agent operates in 1,200 clinics affiliated with Practo, helping patients book visits or follow up on next steps after receiving care. However, calls still account for more than 70% of the company’s business, which has also begun expanding into chat and WhatsApp. For some customers, such as Shell, Ringg automates support requests carried out through the browser.

A coordination layer instead of owning the full stack

Ringg is developing its own speech recognition and generation models, and says it eventually wants to own the entire voice stack, including infrastructure and deployment. For now, however, it operates as a coordination layer that routes tasks to different models depending on the use case, because owning the full infrastructure is expensive at this stage.

The company also does not aim to sell its services directly to U.S. businesses. Instead, it wants to work with global capability centers in India—centers that multinational companies rely on for support and back-office operations—to provide automation alongside human support.

Why does this funding matter?

The importance of the news lies not only in the funding amount, but also in Ringg’s attempt to move from selling minutes or low-complexity automated calls to delivering specific enterprise outcomes. This shift could make the customer relationship and the workflow executed more important than the AI model used behind the scenes.

But India’s market is crowded. Model-building companies such as Deepgram, ElevenLabs and Cartesia, along with Sarvam and Smallest.ai, compete alongside coordination-focused companies such as Bolna and Blue Machines, and others specializing in particular sectors, such as Gnani and Arrowhead. Ringg’s ability to reduce the cost of operating models and maintain consistent quality on complex tasks therefore remains a crucial factor that the funding round alone has not proven.

Ringg currently has 40 employees, more than 15 of whom joined during the past three months. The company employs engineers who work directly with customers and combine technical skills with product management, along with researchers focused on reducing the cost of operating its models.

News source
TechCrunch AI
Open original source ↗
c
Author

certi.news Editorial Team

What you need to know

India-based Ringg has raised $10 million from Peak XV Partners in an extension of its Series A round, bringing the round’s total to $15.5 million. The company uses voice agents to perform enterprise tasks such as scheduling appointments and verifying identities, while expanding into chat, WhatsApp, and browsers.

  • Ringg handles around 20 million call attempts per month, with calls still accounting for more than 70% of its business.
  • The company began as DesiVocal, developing text-to-speech technology, before shifting to building voice agents for enterprises.
  • Ringg’s customers include Cred, Flipkart, Practo, Groww, and Policybazaar, and its voice agent operates across 1,200 Practo clinics.
  • The company is expanding its use cases to appointment scheduling, abandoned-cart recovery, and customer onboarding and identity verification through KYC.
  • Ringg currently operates as an orchestration layer that routes tasks to different models, with ambitions to own the full voice stack in the future.
  • The company faces competition from firms such as Deepgram, ElevenLabs, Sarvam, Smallest.ai, and Bolna, while the funding round alone has not proven its ability to reduce costs and maintain quality.

FAQ

How much funding did Ringg raise?

Ringg raised $10 million from Peak XV Partners, bringing the total size of its Series A round to $15.5 million.

What tasks do Ringg’s voice agents perform?

The agents perform tasks such as outbound calls, lead qualification, and loan collection, as well as clinic appointment scheduling, abandoned-cart recovery, and customer identity verification.

What role do calls play in Ringg’s business?

Calls account for more than 70% of the company’s business, but it has also begun expanding into chat, WhatsApp, and support for tasks carried out through browsers.

Does Ringg own a full voice stack?

Not yet. It currently operates as an orchestration layer that uses different models depending on the use case, while developing its own speech recognition and generation models and aiming to own the full stack in the future.

In the same category

You may also like

View all news