Karta AI raises $3 million seed to automate bank customer calls

Bengaluru voice AI startup Karta AI has raised $3 million (about Rs 28 crore) in a seed round led by PeerCapital and Cedar Hill Capital. Wyser VC, CDM Capital and Seedstars also joined, the company said in a press release on Thursday. It didn't disclose its valuation.

The company builds voice AI agents that handle inbound customer calls for banks, lenders, fintechs and insurers. It plans to spend the money on end-to-end automation, multilingual capabilities, more deployments with financial institutions, and scaling its engineering and go-to-market teams.

Who is behind Karta AI?

Founders Vijaypal Singh and Abhishek Kushwaha launched the startup in 2024 after stints at Razorpay. Singh said the two had worked on internal AI automation at the fintech company before leaving to start Karta AI.

Karta AI started with chat-based automation and moved into voice in August 2025. It now runs an omnichannel platform spanning voice, chat, email and WhatsApp, though voice is the main focus.

The startup has 10-12 paying customers, according to Singh, including PhonePe, Pine Labs, Bajaj and Aavas.

What does the platform do?

Most voice bots handle one query at a time. Karta AI's agents are built to deal with several customer requests, or intents, in a single conversation, and to execute the workflows by connecting with a lender's or insurer's enterprise systems.

A borrower might call to check a loan disbursal status and request a foreclosure statement on the same call. The agent has to recognise both requests, pull information from different systems and start both processes.

The startup targets what the industry calls level 2 and level 3 queries — requests needing data from multiple systems — rather than level 1 queries such as basic FAQs answered with a single lookup.

Singh laid out the economics: level 1 queries make up 30-40% of customer service volumes but only about 10% of costs, while level 2 and level 3 queries together account for around 80% of customer service costs.

Karta AI's proprietary predictive context engine draws on standard operating procedures, knowledge bases, past conversations handled by human agents and AI, and data from enterprise systems. That lets an agent handle follow-up requests in the same call without treating each query in isolation.

The startup uses models from OpenAI and Google's Gemini family, plus open-source models such as Kimi and Gemma. Calls involving fraud or other high-risk matters, and processes without clearly defined standard operating procedures, go to human agents. Customers can also ask to speak to a person.

How crowded is the BFSI voice AI field?

Karta AI's raise extends a run of funding for Indian startups building voice AI for banking, lending and insurance.

• Desible.ai recently raised Rs 32 crore in a Seed Plus round led by Prime Venture Partners.
• Navana.ai raised Rs 40 crore in September in a Series A led by upGrad co-founder Ronnie Screwvala. The company says it has delivered over 100 million voice AI minutes to BFSI clients.
• Rezolv, based in Mumbai, raised $12.5 million in August in a Series A led by Norwest. Its agents work across 11 Indian languages for sales, customer service and collections.

Other players include Arrowhead, which focuses on voice AI for financial services, and Gnani.ai. While some platforms concentrate on specific workflows such as collections and underwriting, Karta AI has bet on inbound calls — customers reaching out about failed payments, blocked cards, loan queries and insurance claims.

With fresh capital and marquee clients like PhonePe and Pine Labs already on board, Karta AI will now test whether multi-intent voice automation can take share in a segment where at least four rivals raised money within months.