₹32cr seed for AI that calls bank and insurance customers

Think of the last time an insurer called to remind you a premium was due. Someone read a script, you said "I'll do it later," and nothing on their side changed until another person called again. Desible.ai, a Bengaluru startup, builds AI agents that make that call and then finish the job: confirm the renewal, send the reminder, update the policy record. On 7 October 2026 it raised ₹32 crore in a Seed Plus round led by Prime Venture Partners.

Who Desible.ai is and What It Sells

Desible.ai sells to BFSI companies (banking, financial services and insurance). That means banks, NBFCs and insurers, plus the TPAs (third-party administrators, firms that handle claims and medical checks on an insurer's behalf).

Uttam Tiwari (growth and fundraising) and Omkar Raikar (tech and product) founded it in 2025. Both have 12+ years of experience, and both co-founded and sold an edtech startup before this one. Raikar previously worked at Crisil and Edelweiss.

The product is what the company calls agentic workflows: AI agents that carry a customer task from start to finish. The company says it runs 25+ of them across sales, collections, underwriting (the insurer's check before it agrees to cover someone), service and claims. Voice is where it started; WhatsApp, SMS and email came next.

What "doing the Job, Not Just the Call" Looks Like

Most phone automation in banking is an IVR, the "press 1 for balance" menu. It answers. It doesn't act. Desible.ai connects its agent to the systems the company already runs, like its telephony, CRM and back-office software, so the conversation can change a record.

Desible.ai's published case study with an IRDAI-licensed TPA shows what that means. Before you get a health policy, the insurer often needs a PPMC (pre-policy medical check-up). The TPA was booking thousands of these a day, and missed appointments were holding up policies. Desible.ai's agent now:

1. Calls the customer the day before and one hour before the check-up to confirm or reschedule
2. Calls again afterwards to check the visit happened
3. Calls the phlebotomist (the person who collects the blood sample) if they are running late, and updates the customer before the customer rings the helpline
4. Closes each case as completed, rescheduled, no-show or escalated, and logs it to the TPA's system

The TPA reports 18,000+ calls a day handled, 89% less manual follow-up and a 33% higher completion rate.

The same pattern runs in the company's other listed use cases: policy renewals, lapse prevention, collections and payment reminders, lead qualification, welcome calls, KYC (identity checks) and claims updates. For a wider view of agents that take over repetitive office work, see how another Indian startup raised ₹4.25 crore to kill copy-paste ops.

site: doableclaw.com — We monitor buyer prompts across AI platforms and deliver clear fixes to close your visibility gaps.

Why Rules Come Before Features in Indian BFSI

A bank or insurer can't let software talk to customers the way a retail app can. In India, four sets of rules apply at once:

• IRDAI, the insurance regulator, sets what can be said to a policyholder
• RBI, the central bank, sets the rules for banks and NBFCs
• TRAI/DLT, the telecom regulator's system, controls who can send commercial calls and SMS
• The DPDP Act 2023, India's data protection law, decides how customer data is stored and used

Desible.ai says its platform is built around all four. On security, its site lists SOC 2 Type II certification (an outside audit that checks security controls work over months, not just on the day of the audit), encryption, single sign-on, audit logs and role-based access, where each staff member sees only what their role needs.

That is why Prime's principal, Gaurav Ranjan, framed the investment around rules rather than voice quality:

The Numbers, and Who Reported Them

Every number below comes from Desible.ai itself, either in its funding announcement or on its website. None are independently audited.

Read them as the company's own claims. The fair test for a buyer is to ask for a reference call with a live client in their own segment.

Who Put in the ₹32 Crore, and What It is For

Prime Venture Partners led the Seed Plus round. Invention Engine, which had already invested, joined again. Reports put the round between $3.3 million and $3.7 million depending on the exchange rate used, so the rupee figure, ₹32 crore, is the one to quote.

The company says the money will go to four things:

1. Deeper agentic AI capabilities
2. Stronger compliance infrastructure
3. A bigger go-to-market push
4. Faster product development across BFSI

Today most of its visible work is in insurance. Both public case studies are insurers or an insurance TPA. The next step it names is the wider BFSI set: banks and NBFCs, on more channels and in more Indian languages. Voice in Indian languages is its own product question; this piece on voice AI agents for Indian languages covers what works.

What This Tells Founders Selling to Banks and Insurers

Pick one task and finish it. Desible.ai's case studies each own one job, like a medical check-up booking or a lead callback, and close it with a logged result.

Name the rules on the first page. IRDAI, RBI, TRAI/DLT and DPDP appear in Desible.ai's own description of itself, not in a footnote. A compliance team reads that before a product team reads the demo.

Show the record, not the recording. The TPA case ends with every appointment closed and logged to the client's system. That log is what a regulated buyer checks.

Indian investors are backing AI at seed across very different buyers. For a contrast, see the ₹15.5 crore bet on AI marketing for local shops.

site: doableclaw.com — Capture leads from AI search when buyers compare. Four assistants, a score per engine, every answer readable.

Conclusion

Desible.ai raised ₹32 crore on the idea that banks and insurers will buy AI agents that finish a regulated task, not agents that only sound good on a call. Its proof so far is insurance work, like medical check-up bookings and renewals, with numbers the company reports itself. The money is meant to take the same approach to banks and NBFCs.