10 AI Voice Scripts for D2C Brands — Copy-Paste Ready for COD, Cart Recovery & Win-Back (2026)

TABLE OF CONTENTS
- Why AI Voice Scripts Are Different From Human Call Scripts
- The Anatomy of a High-Converting AI Voice Script
- Script 1 — COD Order Verification & Confirmation
- Script 2 — COD to Prepaid Conversion
- Script 3 — Abandoned Cart Recovery (High Intent)
- Script 4 — Abandoned Cart Recovery (Medium Intent — Softer Approach)
- Script 5 — Pre-Delivery Day Reminder
- Script 6 — NDR Recovery (Failed Delivery Attempt)
- Script 7 — Post-RTO Win-Back
- Script 8 — Post-Purchase Review Collection
- Script 9 — Reorder / Replenishment Reminder
- Script 10 — Lapsed Customer Win-Back (90+ Days Inactive)
- Universal Best Practices for AI Voice Scripts
- Frequently Asked Questions
BLOG CONTENT
Every D2C founder knows the feeling of watching a COD order disappear into the RTO void.
You shipped in good faith. The courier tried twice. Nobody answered. The order came back. And you are out ₹400–₹600 in logistics, plus the customer acquisition cost you spent to get that order in the first place.
The antidote exists. It is an AI voice call — placed at exactly the right moment, with exactly the right script, asking exactly the right questions to confirm, recover, or re-engage before the loss happens.
But here is the problem most D2C brands face when they try to deploy AI voice calling: they do not know what the agent should actually say.
Generic AI voice platforms give you a blank canvas. You need to write the script. And most scripts written by D2C teams are either too robotic ("Please press 1 to confirm your order"), too salesy ("Act now and get 20% off!"), or too vague to actually move the customer.
This guide solves that problem completely.
Below are 10 battle-tested AI voice scripts — covering every major D2C use case from COD verification to customer win-back — written specifically for Indian D2C brands, optimised for Hinglish-friendly conversational flow, and ready to copy directly into your AI voice platform.
Each script includes: the trigger timing, the target customer, the full opening dialogue, objection-handling branches, the call-to-action, a Hindi/Hinglish variant where relevant, and the expected conversion benchmark.
1. Why AI Voice Scripts Are Different From Human Call Scripts
Before you copy any script, understand one critical difference: AI voice agents are not reading scripts the way a human agent does.
A human agent reads a script linearly and improvises. An AI voice agent follows a conversation tree — a structured flow where each customer response triggers a specific branch. Your "script" is actually a set of responses for every scenario the customer might present.
This means your script must anticipate:
- The customer who confirms immediately
- The customer who has a question before confirming
- The customer who wants to cancel
- The customer who is not sure and needs to be guided
- The customer who is rude or unresponsive
- The customer who does not remember placing the order
Every script below is written with these branches built in. You will see [IF CUSTOMER SAYS X] branches throughout — these are not alternative scripts, they are the objection-handling logic your AI needs.
2. The Anatomy of a High-Converting AI Voice Script
Every script in this guide follows the same 5-part structure:
Part 1 — Warm Opening (5–8 seconds) Introduce the agent with a name (not "AI"), state the brand, and state the reason for calling. Never say "This is an automated message."
Part 2 — Context Anchor (5–10 seconds) Reference something specific — the order number, the product name, the cart item. This establishes legitimacy and prevents the customer from hanging up.
Part 3 — Single Clear Ask (5 seconds) Ask one question. Not two. One. The ask should be as simple as "yes or no" wherever possible.
Part 4 — Objection Branches (10–20 seconds each) Prepare responses for the 3–4 most likely customer objections for each use case.
Part 5 — Close and Next Step (5 seconds) End with a clear action — confirm, send a link, connect to a human, or schedule a callback. Never end a call without a defined next step.
Total target call length: 60–90 seconds.
Calls longer than 90 seconds have sharply lower completion rates. Design every script to resolve within this window.
SCRIPT 1 — COD Order Verification & Confirmation
Use Case: Confirm customer intent and verify delivery address for every COD order before dispatch.
Trigger Timing: Within 10–30 minutes of COD order placement.
Target Customer: Any customer who placed a COD order.
Expected Outcome: 70–80% confirmation rate, 20–30% RTO reduction on verified orders.
English Script
AI: "Hi, am I speaking with [Customer Name]?"
(Customer confirms)
AI: "Hi [Name], this is Priya calling from [Brand Name]. You placed an order with us about [X] minutes ago — Order Number [Order ID] — for [Product Name], worth ₹[Amount], with Cash on Delivery. I am just calling to quickly confirm your order before we dispatch it. Can you confirm that you would like to proceed?"
[IF YES — Customer confirms]: AI: "Perfect! And just to confirm, the delivery address is [Address] — is that correct?" (If address confirmed) AI: "Wonderful. Your order will be dispatched within [X] hours and you will receive a tracking link on WhatsApp. Thank you, [Name], have a great day!" (If address needs change) AI: "No problem at all. Could you share the correct address? I will update it right now before dispatch."
[IF UNSURE — Customer is not sure they want the order]: AI: "That is completely fine. Your order is confirmed and you can always cancel before dispatch by replying CANCEL to our WhatsApp message. Is there anything about the product I can help clarify before you decide?"
[IF CANCEL — Customer wants to cancel]: AI: "Understood, [Name]. I will put a hold on the order. Can I ask — was there something specific that made you reconsider? I want to make sure I can help if it was something like delivery time or product questions."
[IF NO ANSWER — Unanswered after 3 rings]: (System triggers SMS fallback) "Hi [Name], we tried calling to confirm your order #[ID] from [Brand]. Reply YES to confirm or NO to cancel. Order holds for 2 hours."
Hinglish Variant (Tier 2 / Tier 3 Recommended)
AI: "Hello, [Name] ji se baat ho rahi hai?"
(Customer confirms)
AI: "[Name] ji, main Priya bol rahi hoon [Brand Name] se. Aapne abhi humse ek order place kiya — [Product Name] ka, ₹[Amount] mein, Cash on Delivery se. Bas ek minute mein confirm karna tha ki aap order receive karna chahte hain?"
[IF YES]: AI: "Bahut achha! Delivery address [Address] hai — yeh sahi hai na?" (Confirmed) AI: "[Name] ji, aapka order [X] ghante mein dispatch ho jayega aur tracking WhatsApp pe milegi. Dhanyavaad!"
[IF CANCEL]: AI: "Koi baat nahi. Main order hold kar deti hoon. Koi problem thi kya — jaise delivery time ya product ke baare mein koi sawaal?"
Pro Tip: Always use a female voice name (Priya, Kavya, Ananya) for COD verification calls — research consistently shows higher answer rates and more cooperative responses compared to generic or male-named agents.
SCRIPT 2 — COD to Prepaid Conversion
Use Case: Offer an incentive to convert a COD order to prepaid during the verification call — the single highest-ROI move in D2C.
Trigger Timing: Embed this into Script 1, after COD confirmation. Do not make a separate call.
Target Customer: COD customers with order values above ₹800.
Expected Outcome: 12–20% COD-to-prepaid conversion rate, 8–15% additional RTO reduction.
Script (Embedded After COD Confirmation)
(After the customer has confirmed the order)
AI: "[Name], your COD order is confirmed. I also wanted to quickly let you know — if you would like to switch to online payment right now, we will give you an instant ₹[Amount] discount on your order. It takes just 30 seconds and I can send you a payment link on WhatsApp. Would you like that?"
[IF YES]: AI: "Wonderful! I am sending the secure payment link to your WhatsApp number right now. The discount will be applied automatically. Thank you so much, [Name]!"
[IF NO / PREFER COD]: AI: "Absolutely no problem — your COD order is confirmed as is. We will dispatch it shortly. Have a great day!"
[IF MAYBE / WANTS TO THINK]: AI: "Of course! I will send you the prepaid link on WhatsApp anyway — the offer is valid for the next 2 hours if you change your mind. No pressure at all."
Hinglish Variant
(After confirmation)
AI: "[Name] ji, aapka order confirm ho gaya. Ek chhoti si baat — agar aap abhi online payment kar lein, toh hum aapko ₹[Amount] ka instant discount denge. Main abhi WhatsApp pe link bhej deti hoon — 30 second ka kaam hai. Karuun?"
[IF YES]: AI: "Bilkul! Link abhi ja raha hai aapke WhatsApp pe. Discount automatically apply ho jayega. Thank you [Name] ji!"
[IF NO]: AI: "Koi baat nahi! COD order confirmed hai. Jaldi dispatch hoga!"
Pro Tip: Keep the discount amount specific and meaningful — ₹30 feels small, ₹75–₹100 feels significant. For orders above ₹2,000, even a 5% discount (₹100) achieves conversion rates of 18–22%.
SCRIPT 3 — Abandoned Cart Recovery (High Intent)
Use Case: Call a customer who added a high-value product to cart, reached checkout, but did not complete payment.
Trigger Timing: 30–60 minutes after abandonment. Do not wait longer — intent decays rapidly.
Target Customer: High-intent abandoners — visited product page 2+ times, added to cart, reached checkout or payment page.
Expected Outcome: 15–22% cart recovery rate.
Script
AI: "Hi, is this [Customer Name]?"
(Customer confirms)
AI: "Hi [Name], this is Aanya from [Brand Name]. I noticed you were checking out our [Product Name] a little while ago but did not complete the order. I just wanted to reach out personally — was there anything I can help with? Sometimes there are questions about sizing, delivery, or payment that I can sort out right now."
[IF PRICE OBJECTION — "It was too expensive"]: AI: "I completely understand. Just for you, I can offer a [X]% discount if you complete the order in the next hour — I will send you a link with the code already applied on WhatsApp. Would that work?"
[IF DELIVERY CONCERN — "The delivery was too long"]: AI: "Totally valid. Let me check — for your pincode [XXXX], we can actually deliver within [X] days. Does that work better for you?"
[IF PRODUCT DOUBT — "I was not sure about it"]: AI: "I get that. What specifically were you unsure about? I can answer any product questions right now — and we also have a [return policy] if it does not work for you."
[IF JUST BROWSING / NO INTENT]: AI: "No worries at all! I will send you a link to your cart so you can come back whenever you are ready. We have got [X] units left in stock, so I just wanted to make sure you do not miss out."
[IF NO ANSWER]: (Trigger WhatsApp fallback with cart link and a soft time-limited offer)
Hinglish Variant
AI: "[Name] ji, main Aanya bol rahi hoon [Brand] se. Aapne kuch time pehle [Product] checkout karna start kiya tha — koi help chahiye thi kya? Sizing, delivery, ya payment ke baare mein koi sawaal ho toh main abhi solve kar sakti hoon."
[IF PRICE]: AI: "Samajh sakti hoon. Aapke liye main [X]% discount arrange kar sakti hoon — abhi WhatsApp pe link bhejti hoon. Theek lagega?"
Pro Tip: Do not open with "I see you abandoned your cart." That phrasing is off-putting. Instead, frame it as "I wanted to reach out personally" — it positions the call as service, not surveillance.
SCRIPT 4 — Abandoned Cart Recovery (Medium Intent — Softer Approach)
Use Case: Call a customer who added to cart but did not reach checkout — moderate intent, needs a gentler touch.
Trigger Timing: 60–90 minutes after abandonment.
Target Customer: Customers who added to cart but did not initiate checkout. First-time visitors.
Expected Outcome: 8–12% cart recovery rate.
Script
AI: "Hi, is this [Customer Name]?"
(Customer confirms)
AI: "Hi [Name], this is Meera from [Brand Name]. You were browsing our [Product Name / Category] a little while ago and we wanted to make sure you had all the information you need. Is there anything about the product I can help clarify?"
[IF INTERESTED]: AI: "Great! I will send you the product link directly on WhatsApp along with our current offers. It takes 30 seconds to order. Shall I send it now?"
[IF NOT INTERESTED / WRONG TIME]: AI: "Absolutely fine! I will send you the link anyway so you have it when you are ready. We also have a [return policy] in case you want to try it risk-free. Have a lovely day!"
[IF BUSY / CALL BACK LATER]: AI: "Of course! When is a better time? I can schedule a quick callback or just send all the details on WhatsApp — whichever you prefer."
Pro Tip: For medium-intent customers, the goal is not to close on the call. The goal is to move them one step forward — get them to WhatsApp where a chatbot can complete the sale. Keep the call under 60 seconds.
SCRIPT 5 — Pre-Delivery Day Reminder
Use Case: Call the customer the evening before their expected delivery date to ensure availability and reduce first-attempt failure.
Trigger Timing: Evening (6–8 PM) the day before expected delivery.
Target Customer: All customers with shipments in transit — especially COD customers.
Expected Outcome: 15–25% reduction in first-attempt delivery failure.
Script
AI: "Hi [Name], this is a quick call from [Brand Name]. Your order — [Product Name] — is arriving tomorrow! Our delivery partner will attempt delivery between 10 AM and 7 PM. We wanted to make sure someone is available at [Delivery Address] to receive it."
[IF AVAILABLE]: AI: "Perfect! You will also receive a call from our courier before they arrive. Is there a specific time slot in the morning or afternoon that works better for you?"
[IF NOT AVAILABLE TOMORROW]: AI: "No problem at all. Would you like me to request a delivery reschedule? I can try for [Next Day] instead — would that work?"
[IF ADDRESS NEEDS CHANGE]: AI: "Of course — what is the updated address? I will flag it with our delivery team right away."
[IF COD — Remind Amount]: AI: "Just a heads up — this is a Cash on Delivery order, so please keep ₹[Amount] ready. Alternatively, if you would like to pay online now and avoid the hassle, I can send a payment link on WhatsApp."
Hinglish Variant
AI: "[Name] ji, [Brand] ki taraf se ek chhoti si call — aapka [Product] kal deliver hone wala hai! Delivery partner subah 10 baje se sham 7 baje ke beech aayenge. [Address] pe koi available rahega na?"
[IF COD]: AI: "Ek baat aur — yeh COD order hai, toh ₹[Amount] ready rakhiyega. Ya phir agar aap abhi online pay kar lein toh main link bhej sakti hoon — aapki marzi."
Pro Tip: This call consistently delivers the highest ROI-to-effort ratio of any voice use case. It costs one call to prevent a potentially ₹600 RTO. Always make this call for COD orders above ₹500.
SCRIPT 6 — NDR Recovery (Failed Delivery Attempt)
Use Case: Customer was not available when the courier arrived. Recover the delivery before it becomes an RTO.
Trigger Timing: Within 30–60 minutes of receiving an NDR (Non-Delivery Report) from the courier.
Target Customer: Any customer with a failed delivery attempt.
Expected Outcome: 30–45% of NDRs recovered through voice + WhatsApp cascade.
Script
AI: "Hi [Name], this is a call from [Brand Name]. Our delivery partner just tried to deliver your [Product Name] order today but could not reach you. We want to make sure you get your order — can we help arrange a redelivery?"
[IF CUSTOMER WAS JUST BUSY]: AI: "Completely understand! Our courier can attempt delivery again tomorrow. Is morning or afternoon better for you at [Address]?"
[IF WRONG ADDRESS / NEEDS CHANGE]: AI: "Let me note that down right now. What is the updated address or any landmark that can help the courier find you?"
[IF WANTS TO CANCEL]: AI: "I am sorry to hear that. Before I process the cancellation — was there something about the product or the delivery experience I can help with? If it is a time issue, we can hold the order for up to 3 more days."
[IF COD — Offer Prepaid]: AI: "Also, if it helps — you can pay online now and we will mark the order as priority for tomorrow's delivery. I can send a payment link right away on WhatsApp."
[IF NO ANSWER — Missed the AI Call]: (Trigger WhatsApp message) "Hi [Name], we tried to reach you — your [Product] delivery was attempted today. Reply 1 to reschedule for tomorrow, 2 to change address, or 3 to cancel. We will hold your order for 24 hours."
Hinglish Variant
AI: "[Name] ji, [Brand] se call aa rahi hai. Aaj aapka [Product Name] deliver karne ki koshish ki gayi thi lekin aap available nahi the. Hum chahte hain ki aapko order mile — kya kal dobara delivery arrange kar sakte hain?"
[IF YES]: AI: "Bilkul! Kal subah ya dopahar — kaun sa time better rahega? [Address] pe deliver karein na?"
Pro Tip: Speed is everything for NDR recovery. An NDR that goes unaddressed for 24 hours has a 70%+ chance of becoming an RTO. Set your NDR voice trigger to fire within 30 minutes of the NDR report arriving from your courier partner.
SCRIPT 7 — Post-RTO Win-Back
Use Case: The order has already been returned. One last attempt to recover the customer and re-convert the sale.
Trigger Timing: Within 24 hours of the RTO being logged in your system.
Target Customer: Customers whose COD order was returned after 2+ failed delivery attempts.
Expected Outcome: 10–18% win-back rate (orders that successfully re-convert).
Script
AI: "Hi [Name], this is Riya from [Brand Name]. I am calling about your recent order — [Product Name] — which was unfortunately returned to us as we were unable to complete the delivery. I just wanted to personally check in and see if you would still like to receive the order."
[IF YES — Wants to Reorder]: AI: "Wonderful! I can place the order again right now. And just as a token of apology for the inconvenience, I will add a ₹[Amount] discount on the reorder. Would you prefer COD again or shall I send a payment link?"
[IF NO — Changed Mind]: AI: "Completely understand, [Name]. Can I ask — was it something about the product, or were there other reasons? Your feedback genuinely helps us improve. And if you would like to give us another shot in the future, I will make sure we have a special offer ready for you."
[IF ADDRESS ISSUE WAS THE PROBLEM]: AI: "I see — so the delivery address was the issue. If you give me the correct address right now, I can re-dispatch the order today with priority delivery. The product is still here with us."
[IF BUSY / NOT THE RIGHT TIME]: AI: "Absolutely fine. I will send all the details on WhatsApp, including a special reorder link. It will be valid for 48 hours — no pressure at all."
Pro Tip: Do NOT start this call with "your order was returned." Frame it as a personal outreach to check in — "I wanted to personally ensure you get your order." The customer is already frustrated; empathy-first language significantly improves receptivity.
SCRIPT 8 — Post-Purchase Review Collection
Use Case: Collect a product review or NPS rating from the customer 2–3 days after confirmed delivery.
Trigger Timing: 2–3 days after delivery confirmation from the courier.
Target Customer: All successfully delivered customers, prioritising first-time buyers.
Expected Outcome: 30–45% review collection rate (vs 5–10% for email review requests).
Script
AI: "Hi [Name], this is a quick call from [Brand Name]. We noticed your [Product Name] was delivered a couple of days ago — we just wanted to check in and see how you are enjoying it?"
[IF POSITIVE RESPONSE]: AI: "That is wonderful to hear! Would you be happy to leave us a quick review? It takes less than a minute and really helps other customers like you. I can send the review link to your WhatsApp right now — would that be okay?"
[IF NEGATIVE / ISSUE WITH PRODUCT]: AI: "I am really sorry to hear that, [Name]. Could you tell me a little more about what happened? I want to make sure we sort this out for you right away — whether it is an exchange, a return, or something else."
[IF NEUTRAL / STILL TRYING]: AI: "Great! Give it a few more days and we hope you love it. Once you have had a chance to try it properly, we would love your feedback. I will send you a review link on WhatsApp — no rush at all."
Pro Tip: Ask one open question first ("How are you enjoying it?") before asking for the review. This primes positive sentiment and dramatically increases the rate of 4–5 star reviews compared to leading directly with the review request.
SCRIPT 9 — Reorder / Replenishment Reminder
Use Case: Call a customer who previously purchased a consumable product (supplements, skincare, food, personal care) when they are likely to be running low.
Trigger Timing: Based on product consumption cycle — typically 25–30 days after delivery for most consumables.
Target Customer: Repeat-purchase customers with a confirmed delivery on consumable products.
Expected Outcome: 20–35% reorder conversion rate.
Script
AI: "Hi [Name], this is a call from [Brand Name]. You ordered [Product Name] about a month ago — and most customers find they start running low around this time. I just wanted to check if you would like to reorder before you run out?"
[IF YES]: AI: "Perfect timing! I can get your reorder placed right now with the same delivery address — [Address]. And as a loyal customer, you also get [X% off / free shipping] on this reorder. Should I go ahead?"
[IF STILL HAS STOCK]: AI: "No worries at all! I will make a note to call you again in [X] days. And whenever you are ready, you can also reorder directly on our website or just reply REORDER to our WhatsApp."
[IF SWITCHED BRANDS]: AI: "I completely understand. Can I ask what made you switch? We are always looking to improve — and if it was something about the product, I might be able to help."
[IF LOVED THE PRODUCT]: AI: "That is fantastic! While I have you — we also have [Complementary Product] that works really well alongside [Product Name]. A lot of our customers who love [Product] also love [Complementary]. Would you like me to send details on WhatsApp?"
Hinglish Variant
AI: "[Name] ji, [Brand] se bol rahi hoon. Aapne [Product] karib ek mahine pehle order kiya tha — aur generally is time tak log low pe aa jaate hain. Kya aap reorder karna chahenge?"
Pro Tip: This is the highest-LTV call in your entire D2C voice stack. A successful reorder call costs ₹6–₹12 and generates 1.5–3x the AOV of the first purchase (because repeat buyers convert at higher rates and often add complementary products). Run this call for every consumable category every purchase cycle without exception.
SCRIPT 10 — Lapsed Customer Win-Back (90+ Days Inactive)
Use Case: Re-engage a customer who purchased once but has not returned in 90+ days.
Trigger Timing: Day 90 after last purchase (or after no engagement with WhatsApp/email for 60 days).
Target Customer: One-time buyers who have not repurchased. Specifically target those whose first purchase was a positive experience (no returns, no complaints).
Expected Outcome: 10–18% win-back rate.
Script
AI: "Hi [Name], am I speaking with you?"
(Customer confirms)
AI: "Hi [Name], this is a call from [Brand Name]. You ordered [Product Name] from us a few months ago and we have genuinely missed having you. We wanted to reach out personally with something special — would you have just a minute?"
[IF YES]: AI: "We have got a [X% off / special welcome-back offer] just for you — valid only this week. And we have also launched some exciting new products since your last order that I think you will love. Shall I send the details along with your exclusive offer to your WhatsApp?"
[IF PRICE SENSITIVE]: AI: "Absolutely understood. The welcome-back offer we have for you is actually our best discount of the year — [X%] off with free delivery. I will send it on WhatsApp so you have it whenever you are ready."
[IF BAD FIRST EXPERIENCE]: AI: "I am really sorry to hear that, [Name]. That is exactly why I am calling — I want to make it right. Can you tell me what went wrong? I would like to personally ensure your next order is a completely different experience."
[IF NOT INTERESTED]: AI: "No worries at all, [Name]. I appreciate you taking the time to speak with me. If you ever change your mind, we will always have a special offer ready for you. Have a wonderful day!"
Pro Tip: Never open a win-back call with "We noticed you haven't ordered in a while." It sounds passive-aggressive. Use "We have genuinely missed having you" — it is warm, human, and positions the brand as one that values the relationship, not one that is chasing revenue.
3. Universal Best Practices for AI Voice Scripts
1. Always Use a Named Agent, Not "AI" or "Automated System"
Your AI agent should always have a human name — Priya, Aanya, Meera, Riya — and should introduce itself without mentioning that it is AI or automated. Modern AI voices are indistinguishable from humans, and disclosing "this is an automated call" at the start causes immediate hang-ups in India.
2. Time Your Calls Correctly
| Use Case Ideal Calling Window | |
| COD Verification | 10–30 min after order |
| Abandoned Cart | 30–90 min after abandonment |
| Pre-Delivery Reminder | 6–8 PM the evening before |
| NDR Recovery | Within 30–60 min of NDR report |
| Post-RTO Win-Back | Within 24 hours of RTO logged |
| Review Collection | 2–3 days post-delivery |
| Reorder Reminder | Day 25–30 of consumption cycle |
| Lapsed Win-Back | Day 90 post last purchase |
3. Never Call Between 9 PM and 9 AM
TRAI regulations in India prohibit commercial calls between 9 PM and 9 AM. Beyond compliance, calls during these hours have significantly lower answer rates and generate negative brand sentiment. Schedule all AI voice triggers to queue and fire only between 9 AM and 8:30 PM.
4. Always Have a WhatsApp Fallback
Every script above that mentions "I will send you a link on WhatsApp" needs that link to actually fire automatically the moment the agent says it. Your AI voice platform should be integrated with your WhatsApp automation so that the promised link, offer, or information reaches the customer within seconds of the call ending — whether the call was answered or not.
5. Keep Objection Branches Tight
Each objection branch should be 1–2 sentences maximum. Long objection responses lose customers. The goal of handling an objection is not to win an argument — it is to remove one specific barrier and move to the next step.
6. Log Every Call Outcome
Ensure your AI platform logs structured outcomes for every call: Confirmed / Cancelled / Rescheduled / No Answer / Objection Type. This data feeds your intent scoring model and improves every subsequent campaign.
7. A/B Test Your Openings
The opening 8 seconds of a call determine whether the customer stays or hangs up. Test at least two opening variants per use case — for example, mentioning the product name first vs. the brand name first. Even small changes can move answer-to-engagement rates by 15–25%.
4. Frequently Asked Questions
Can I use these scripts directly in Retner's AI voice agent?
Yes. These scripts are structured to match Retner's conversation tree format. You can paste each script section into the corresponding node — opening, confirmation, objection branches, and close. Retner's platform handles the voice synthesis, telephony, timing trigger, and WhatsApp fallback automatically.
How do I choose between English and Hinglish for my AI voice script?
A simple rule: use Hinglish for customers in Tier 2 and Tier 3 cities (based on delivery pincode), and for customers whose order patterns suggest lower digital engagement. Use English for metros and for customers who have previously engaged with English-language brand communication. Retner's system can auto-route to the right language variant based on customer location data.
Should my AI agent reveal it is AI?
This is a nuanced question. Retner's recommendation is to use a natural-sounding human name and voice without proactively announcing "this is an AI." However, if a customer directly asks "Am I talking to a human or a bot?", the agent should answer honestly. Deceptive practices around AI identity create trust issues and have emerging regulatory implications under India's DPDP Act.
How long should an AI voice call be?
The sweet spot for D2C AI voice calls is 60–90 seconds. Calls under 45 seconds feel rushed and incomplete. Calls over 2 minutes have significantly lower completion rates — customers disengage. Design every script to resolve within 90 seconds for a confirmed customer, with a natural exit for non-responders.
What happens if the customer does not answer?
All 10 scripts above trigger a WhatsApp fallback automatically when a call goes unanswered. This ensures every customer gets at least two touchpoints regardless of call answer rate. For COD verification and NDR recovery, an SMS fallback fires simultaneously with the WhatsApp message.
How many retry attempts should the AI make if the customer does not answer?
For COD verification and NDR recovery: 2 attempts (first attempt + one retry 2–3 hours later). For cart recovery: 1 attempt only — a second unanswered cart recovery call feels intrusive. For all other use cases: 1 attempt + fallback to WhatsApp.
Do these scripts work for fashion, beauty, and supplements differently?
The structure is the same across categories, but personalise the objection branches. Fashion customers often have size doubts — add a size guide offer in your objection branch. Beauty customers need safety reassurance — add "our products are dermatologist tested" or "we have a [X]-day return policy." Supplement customers respond to clinical evidence mentions — reference certifications or key ingredients in your product doubt branch.
Ready to deploy these scripts on your Shopify store? See how Retner's AI Voice Engine works — Book a free demo →
8x Return On WhatsApp Ads Spend.
Book a demo and we will get you started on proven WhatsApp marketing strategies