Pomegra
Karta AI, a Bengaluru voice AI startup founded in 2024, has raised a $3 million seed round to automate complex inbound bank and insurer calls in India.
Key Takeaways
- Karta AI raised a $3M seed led by PeerCapital and Cedar Hill Capital; valuation was not disclosed.
- Its agents resolve about 70% of borrower queries end to end across 25+ languages.
- The startup handles roughly 2 million customer conversations a month.
Lead
Karta AI, a Bengaluru-based startup building voice agents for banks, lenders, fintechs and insurers, announced a $3 million (about ₹28 crore) seed round on October 22, 2026. PeerCapital and Cedar Hill Capital led the round, with Wyser VC, CDM Capital and Seedstars participating. The company did not disclose its valuation or any earlier funding.
Founded in 2024 by Vijaypal Singh and Abhishek Kushwaha, both former Razorpay employees who worked on internal AI automation, Karta targets the calls that conventional voice bots tend to drop. It says it handles about 2 million customer conversations a month and resolves roughly 70% of borrower queries without human intervention.
What Does Karta AI Actually Build?
Karta AI builds an omnichannel customer communication platform spanning voice, chat, email and WhatsApp, with voice as the main product. The pitch centers on multi-intent conversations, where a single caller needs several answers or several back-end systems touched in one call.
Typical cases include loan disbursal status, account statements, foreclosure statements and property document release. Others include failed payments, blocked cards and insurance claims. A "predictive context engine" draws on a customer's earlier interactions, and when a call does go to a human, the conversation history goes with it.
The platform supports more than 25 languages, including Hindi-English code-switching, which matters in a market where callers often change language mid-sentence. The company says its agents cut cost per query by about half.
How Much Traction Does the Startup Have?
Karta AI has between 10 and 12 paying customers, a modest base for a company whose headline volume is 2 million conversations a month. Customers named in coverage of the round include PhonePe, Pine Labs, Bajaj and Aavas. The company reports 70% to 80% automation rates in proof-of-concept deployments.
Those figures come from the company itself and have not been independently verified. The 70% resolution rate also applies to one category, borrower queries, rather than to every call type the platform handles. Production performance across a wider set of banking and insurance workflows is the number to watch.
Why Does a $3M Seed Matter in Indian Voice AI?
The round matters because it points to investor appetite for narrow, regulated-industry voice automation over general-purpose assistants. Banking, financial services and insurance (BFSI) is a high-volume call center market, and the economics are simple: each call resolved without a human removes a per-minute labor cost. Indian lenders also handle large multilingual customer bases, which raises the bar for any bot.
A $3 million seed is small by global AI standards, and the absence of a disclosed valuation limits what can be read into the pricing. Co-leads are common at this stage and spread both the check and the diligence. The mix of investors, which includes an emerging-markets network in Seedstars, suggests a focus on regional distribution rather than a large initial war chest.
Competition is real. Established conversational AI vendors, in-house bots at large banks and other Indian voice AI startups all target the same call volumes. Karta's differentiator, resolving multi-step requests across several enterprise systems, is also the hardest part to deliver reliably, since each integration with a lender's core systems adds engineering and compliance work.
How Will Karta AI Use the Money?
Karta AI will spend the money on three areas: go-to-market expansion over the next two years, wider distribution, and product development. The company singled out self-learning capabilities, meaning agents that improve from resolved and escalated calls, as a development priority. It also plans to strengthen end-to-end automation and multilingual coverage, and to grow its engineering and sales teams. Headcount was not disclosed.
Singh framed the product around difficult calls. "We built Karta AI to resolve the hard calls, the ones that need several systems and more than one answer, and to bring in a human only when it matters," he said.
Outlook
Karta AI enters the market with a clear niche, a Razorpay-rooted team and a customer list that includes recognizable fintech names. The unresolved questions are how far the 70% resolution rate holds as the customer base grows beyond a dozen accounts, and how quickly the startup can convert pilots into paid deployments. A larger follow-on round would depend on those numbers, along with the cost of integrating with each lender's systems.
