Places callsGraph agentsaleslead qualificationloans

Loan lead qualification

Calls people who enquired about a loan, checks basic eligibility and books a callback from an advisor.

How it opens

The first line, in each language the template ships with. The agent answers in whatever language the caller uses.

Hindi
नमस्ते, मैं ‹your company› की एआई सहायक बोल रहा हूँ। आपने लोन के बारे में पूछताछ की थी, क्या दो मिनट बात हो सकती है?
English
Hello, this is an AI assistant from ‹your company›. You enquired about a loan; do you have two minutes?
Hinglish
Namaste, main ‹your company› ka AI assistant hoon. Aapne loan ke baare mein enquiry ki thi, do minute baat ho sakti hai?

The steps

A financial product: the questions and the 'not an approval' line come in a fixed order.

  1. 1

    Greet and ask for two minutes

    Disclose the AI assistant, mention their enquiry, ask if now is a good time.

  2. 2

    What do they need?

    Ask which loan and roughly how much, in rupees.

  3. 3

    Basic profile

    Ask employment type (salaried, self-employed, business), monthly income band and city. Ask one question at a time.

  4. 4

    Offer an advisor

    Say an advisor will explain offers, read the 'not an approval' line and ask for a callback time or offer to connect now.

  5. 5

    Not eligible today

    Thank them kindly, say the current products may not fit, and close without advice.

  6. 6

    Not interested

    Thank them, mark do not call again.

  7. 7

    Callback

    Take a convenient time.

  8. 8

    Hand over

    Transfer to an advisor.

  9. 9

    Close

    Confirm the callback time and thank them.

What each call returns

Filled in after the call, next to the transcript, summary and cost. Delivered by webhook and readable through the API.

FieldQuestion it answers
lead_statusHow did the lead qualify?qualified · not_eligible · not_interested · callback · transferred
loan_typeWhich loan do they want?
loan_amountHow much do they need, in rupees?
employment_typeEmployment type?salaried · self_employed · business · other
monthly_income_bandWhich monthly income band did they state?
cityWhich city are they in?
callback_timeWhen should the advisor call?
do_not_callDid the person ask not to be called again?yes · no

Your systems

No API yet? Keep the lead in extraction only; the customer exports qualified leads from call history.

  • POST save_lead

    Save the qualified lead for the advisor team.

  • TRANSFER transfer_to_human

    Transfer the call to the handover number. Leave it out when no number was given.

Knowledge base: the public product pages (eligibility, documents needed) for sourced answers.

Rules it keeps

Enforced by the platform's compliance toolkit, not left to the script.

AI disclosure
“This call is from an AI assistant of the company, about your loan enquiry.”
Calling window
10:00 to 19:00 in the recipient's timezone. Calls outside it move to the next start.
Transcripts
Masking: card, Aadhaar, PAN and account numbers are hidden before text is stored or sent on, keeping only the last four digits.
Do not call
If the person asks not to be called again, confirm it once, record do_not_call = yes and end the call politely. Tell the customer to add such numbers to the account's do-not-call list.

What you fill in

The form in Set up for me. Most fields have sensible defaults; what you leave empty, the plan marks clearly.

Your business

Company name, Loan products offered, Minimum monthly income (₹)

Agent persona

Agent name, Languages, Voice, Tone

Calling rules

Call from, Call until, Timezone, Human handover number

Your systems

Your API base URL

Script and rules

What to collect, Lines that must be said, Never say or do

Lines to test it with

Type them into the test chat or a workflow test before any real call is placed.

  • “Haan boliye, mujhe personal loan chahiye.”
  • “Paanch lakh ke aas paas, main salaried hoon.”
  • “Interest rate kitna hoga?”
  • “Mujhe interest nahi hai, call mat kijiye.”

Also for banking and cards