A borrower who has missed an EMI by three days is not a defaulter — they are a person who, statistically, has either forgotten, is experiencing a temporary cash flow delay, or is waiting for a salary credit that is two days late. The Early Bucket Caller AI speaks to them like that: as a reminder service, not a collections operation. The script design makes that distinction audible in every interaction.
Why tone is the most important variable in DPD 0–30 collections
In the DPD 0–30 bucket, the institution holds a significant advantage: the borrower's relationship with the lender is still primarily defined by the onboarding experience, not by a collections interaction. Most borrowers in early delinquency have not yet spoken to a collections agent. They do not yet think of themselves as delinquent borrowers. The first contact in this bucket sets the frame for how the borrower will engage with the institution's collections process — and a frame set correctly in the first interaction is far easier to maintain than one that needs to be corrected after a hostile early contact.
Research on early-bucket collections in Indian lending consistently shows that the single highest predictor of resolution in DPD 0–30 is whether the borrower engaged with the first contact. An engagement rate of 70%+ on the first reminder contact produces a promise-to-pay rate that makes formal collections intervention rarely necessary for this bucket. An engagement rate of 30% produces the opposite — the borrower goes silent, the DPD age increases, and the cost of collection escalates rapidly.
The Early Bucket Caller AI's script design is built around maximising that first engagement — with a tone that is warm and helpful, not transactional and threatening.
The four script principles — and why each exists
State who you are before you state what you want
Most collections scripts open with the amount due. This immediately places the conversation in an adversarial frame — the institution wants something from the borrower. The Early Bucket AI opens by identifying the institution, confirming the borrower's identity, and establishing that the call is a service interaction ("we noticed something on your account") before the amount is mentioned. The amount is introduced as information, not a demand.
Default framing assumes intent. Oversight framing assumes the most likely reality.
In the DPD 0–7 window, the most common cause of a missed payment is genuinely forgetting, a salary delay, or a NACH failure the borrower is unaware of. The Early Bucket AI's script assumes the borrower did not intend to miss the payment — because statistically, that is almost always true at this DPD stage. Phrases like "wanted to check if there was a technical issue with the payment" open the conversation on a basis the borrower can engage with honestly, rather than feeling accused.
Use the borrower's name, their specific amount, their specific account.
Generic reminder messages — "your EMI payment is overdue" — feel like mass communications, because they are. The Early Bucket AI personalises every contact with the borrower's name (as they gave it at onboarding, not as it appears in the system — "Rajesh" not "MR R KUMAR"), the precise amount due, the product name they applied for, and the specific date. This specificity signals that the institution is paying attention, which creates a reciprocal engagement obligation in the borrower's perception.
End every interaction with a path to resolution, not a deadline threat.
Even when consequences exist (late payment charges, credit bureau reporting), the DPD 0–30 script ends with an offer — "the easiest way to clear this is..." — not a threat. The credit bureau and penalty information is disclosed because it must be (RBI Fair Practices Code requirement) but it is disclosed as information the institution wants the borrower to have, not as leverage. The Early Bucket AI closes every call with a specific payment link, UPI ID, or branch alternative — removing every friction from the path to payment.
The scripts in practice: Day 1, Day 7, Day 20
Day 1
The missed payment may not even be known to the borrower yet
At DPD 1–3, the first contact is a notification, not a collections call. The AI's message informs — "your EMI wasn't received" — and provides the payment path immediately. No escalation language. No mention of consequences. If the borrower pays within 24 hours of the first message (which 48% do at this stage), the interaction has no negative register in the borrower's experience at all — it felt like a service notification.
Week 1
First voice contact — still entirely soft, but with a gentle time awareness
By DPD 7, the AI escalates to a voice call if the WhatsApp message was not engaged with. The tone remains warm but introduces a mild time dimension — "just wanted to check in before this moves further along" — without specifying consequences. A promise-to-pay is sought at this contact: "Is there a date this week that works for you to make the payment?" This is the most important contact in the DPD 0–30 journey: the conversation that either captures a commitment or reveals a reason the institution needs to know about.
Week 3
Consequences must be disclosed — but as information, not threat
At DPD 20, the RBI Fair Practices Code and credit bureau reporting timelines mean the AI must disclose that the account will be reported to credit bureaus at DPD 30. This is disclosed factually: "I want to make sure you're aware that at 30 days past due, this is reported to credit bureaus — we'd much rather help you resolve this before that happens." The framing is protective, not punitive. A part-payment option is introduced at this stage if the borrower indicates inability to pay the full amount.
Pre-bucket
Last-chance Early Bucket contact before 30+ DPD transition
At DPD 28–30, the Early Bucket AI makes a final contact with full consequence disclosure and a clear statement of the escalation path — "this account will move to our collections team after the 30th." A human agent is pre-queued for this borrower if the AI contact does not produce a payment or commitment. The AI's final contact includes a one-time settlement offer or restructure option where the institution's policy permits it — giving the borrower a meaningful reason to engage before the bucket transition.
A full DPD 7 script — warm, personalised, outcome-focused
The tone of the first contact determines the outcome of the last one
A borrower who feels accused in the first reminder contact becomes a borrower who avoids calls. A borrower who feels helped in the first reminder contact becomes a borrower who answers calls. In the DPD 0–30 bucket, the difference between these two outcomes is almost entirely determined by the first conversation — not by the amount owed, not by the borrower's intent, and not by the institution's collections capacity. The Early Bucket Caller AI's script design is a precision instrument: it is calibrated to produce engagement, and engagement in the early bucket is the variable that matters most.