Lending & Fintech

AI calling for loan recovery

Call borrowers about overdue EMIs, record their payment date and reason, and schedule the next step.

The problem

Collection teams spend hours trying numbers that never answer. When someone does pick up, the follow-up depends on what the agent remembers to record. Missed promises can stay unnoticed until the next collection cycle.

How Vaanicall runs it

  1. Import borrowers

    Upload your overdue-account list as a CSV — that is the production batch-import path today. Every number is automatically scrubbed against the DND registry before a single dial. A contact-ingest API endpoint is on the roadmap.

  2. AI calls in the TRAI window

    The scheduler dials only between 09:00–21:00 IST and honours the DND registry. The opening line is yours to script — and if your licence requires it, optional in-call AI-disclosure enforcement can be switched on per tenant.

  3. Capture promise-to-pay date and reason

    An extraction schema pulls the committed payment date and the stated reason (cash-flow, salary delay, dispute) from the conversation and writes them as structured fields to the contact record.

  4. Retry and callback per disposition

    No-answers re-enter the retry queue at your configured interval. A "promise given" disposition schedules a reminder call before the due date. A "broken promise" escalates to a human agent automatically.

  5. Escalate broken promises

    If a committed date passes without payment, the system raises a callback task for your collection supervisor — no manual queue monitoring required.

Hear it

Sample call — the agent opens with your brand and script

  1. AI Agent: Namaste, main Karan bol raha hoon Arthveer Finance ki taraf se. Kya main Priya Sharma ji se baat kar sakta hoon?

  2. Customer: Haan, main hi hoon. Kya hua?

  3. AI Agent: Priya ji, aapki ₹8,400 ki EMI 12 June ko due thi aur abhi pending hai. Kya aap mujhe bata sakti hain kab payment ho sakti hai?

  4. Customer: Salary thodi late aayi hai. Main 20 tarikh tak kar dungi.

  5. AI Agent: Shukriya Priya ji. Maine 20 June ki promise-to-pay note kar li hai. Kya koi aur date discuss karni hai?

  6. Customer: Nahi, 20 ko ho jaayega.

Sample transcript for illustration. Actual call content varies by contact data and agent configuration.

What you get

Compliance

Compliance

  • TRAI calling window: dials only 09:00–21:00 IST; no out-of-window leaks.
  • DND scrub: numbers on the DND registry are skipped automatically at dial time.
  • Consent record: consent action and timestamp are captured and stored in the audit log.
  • Audit log: every call action is written to an immutable append-only log.

Questions, answered

The platform enforces TRAI calling windows (09:00–21:00 IST) and skips DND-registered numbers automatically. The immutable audit log captures call time, extraction outcome, and retry action for regulator review. In-call AI-disclosure enforcement is available as an optional per-tenant setting if your licence requires it. Your organisation remains responsible for aligning messaging to any RBI-specific guidance for your licence type.

Related use cases

Make your first AI call today

Set up an agent, import contacts, and see what happened on each call. You can start free and pay for connected talk-minutes.