Guides, Chatbot

How Courier and Delivery Companies Use Chatbots to Cut Support Calls

The 10 calls a courier chatbot can take, the 5 it must never handle alone, and the scripts, knowledge base and 14-day pilot plan to cut routine support calls.

L
Laxman
November 16, 2026•13 min read
How Courier and Delivery Companies Use Chatbots to Cut Support Calls
#ai, chatbot

It’s 11:40 a.m. at a courier hub. Four phones are ringing. The supervisor can guess who’s calling before she picks up.

One person wants to know where their parcel is. Another wants the rate to Jaipur. A third was “home all day” and nobody came. The fourth is asking whether you can pick up a 12-kilo box tomorrow.

None of these calls is hard. All of them need a minute or two. And while those four are on the line, a customer with a genuinely damaged shipment is stuck in the queue, getting angrier by the minute.

That’s the real case for a chatbot in a courier business: not to replace your team, but to keep the easy calls from blocking the important ones. This guide goes through it call by call: which ones a chatbot can take, what it should say, where it must hand over, and how to test the whole thing in two weeks.

Start by listening to your own calls

Before you set anything up, spend a day or two noting why people call. Tally the last 100 calls and messages into piles: status questions, rates and serviceability, pickups, missed deliveries, changes, claims, business enquiries and everything else.

You’ll often see claims that “where is my order” (WISMO) contacts make up a fixed share of support volume. Be careful with those numbers. Tracking company Ship24 notes that common WISMO percentages lack a primary source and a clear definition of what counts as a contact, so your own tally is the only figure worth trusting.

The other number worth knowing is what a call costs you. Gartner’s benchmark puts the median cost per contact at about $13.50 for assisted channels such as phone, chat and email, against about $1.84 for self-service. That’s a US-led benchmark and your wage costs will differ, but the gap between a human conversation and a self-service answer is large everywhere. Work out your own: monthly support spend divided by contacts handled.

The 10 calls a courier chatbot can take

For each call below you’ll find what the customer asks, a sample answer, and what the bot needs in its knowledge base. The sample answers are illustrations. Replace the details with your own.

1. “How do I track my parcel?”

Sample answer: “You can track your parcel using the consignment number on your booking receipt or SMS. Enter it on our tracking page: [link]. If you can’t find the number, share the phone number used for booking and our team will help.”

Needs: your tracking page link, where the consignment number appears, and a fallback.

Honest limit: unless your chatbot is connected to your tracking system, it can explain where to look but can’t read a specific parcel’s live status.

2. “What does this tracking status mean?”

Sample answer: “‘Out for delivery’ means your parcel has left the local hub with the delivery partner and should arrive today. ‘Attempted’ means a delivery attempt was made but couldn’t be completed. We usually try again the next working day. You can also ask for redelivery here: [link].”

Needs: a plain-English glossary of every status you use. This single entry removes a surprising number of repeat calls.

3. “What will it cost?”

Sample answer: “Rates depend on weight, size and destination. For parcels up to 500 grams within the same city, the charge starts at ₹[amount]. Tell me the weight and the destination pincode, and I can share the estimate or connect you with our team.”

Needs: your rate card in simple sentences, plus what’s extra (cash on delivery, insurance, remote areas).

4. “Do you deliver to my area, and how long will it take?”

Sample answer: “Metro-to-metro deliveries usually take 2 to 3 working days. Other cities take 4 to 6, and remote locations may take longer. You can check serviceability for your pincode here: [link].”

Needs: typical transit times by zone, and where to check pincodes. Avoid guarantees. “Usually” is honest, and “guaranteed” is a risk.

5. “How do I book a pickup?”

Sample answer: “You can book a pickup on our website or at your nearest branch. Pickups booked before [time] are usually collected the same day. Please have the parcel packed and the address and a contact number ready.”

Needs: pickup cut-off times, how to book, what to have ready and what happens if nobody’s home.

6. “Nobody came, but I was home.”

This is the call that causes the most anger, so the bot’s tone matters most. Sample answer: “I’m sorry about the missed delivery. After a failed attempt, we usually try again within [x] working days. To arrange redelivery, share your name, phone number and a preferred time, and our team will confirm.”

Needs: your redelivery policy, number of attempts and a handoff to a person. The bot should apologise, collect details and promise only what you can deliver.

7. “Can I change the address or timing?”

Sample answer: “Address changes depend on whether the parcel has already left the hub. Share your details and the change you need, and our team will confirm whether it’s possible.”

Needs: your rules on address changes, redirection and delivery windows. The bot shouldn’t promise the change. It should collect the request.

8. “What can’t I send, and how should I pack it?”

Sample answer: “We can’t carry [list of prohibited items]. For fragile items, please use sturdy packaging with padding, and mark the box fragile. Our packaging guide is here: [link].”

Needs: your prohibited and restricted items list, and packaging guidelines. This one also prevents damage claims later.

9. “My parcel is damaged. How do I make a claim?”

Sample answer: “I’m sorry to hear that. To start a claim, we’ll need your consignment number, photos of the parcel and its packaging, and a short description. Share your name and phone number and our claims team will contact you within [time].”

Needs: the claims process, deadlines, documents required and a handoff to a named person.

10. “I run a business. Can you give me a rate or pickup contract?”

These are sales leads, not support calls. Sample answer: “We work with sellers and small businesses on regular pickups and discounted rates. Tell me roughly how many parcels you ship per week and the cities you send to, and our sales team will get in touch.”

Needs: a short description of your business services, and someone who follows up quickly. This is where a chatbot can bring in revenue, not just save time.

The 5 calls a chatbot should never handle alone

Some conversations need a person, and handing them over quickly is part of good design.

  1. Parcels that are genuinely lost. One support-software write-up warns that aggressive automation of status answers can strand the small share of customers whose parcel really is lost, which is exactly when a wrong answer costs the relationship.

  2. Angry, distressed or abusive customers. The bot can acknowledge and collect details. A person should call.

  3. High-value, legal or insurance disputes.

  4. Refunds, compensation and payment disputes. The bot must not offer or promise either.

  5. Anything that needs the live status of a specific shipment, unless your chatbot is connected to your tracking data.

For all five, the bot should say honestly that a person will take over, collect a name and phone number, and tell the customer what happens next (“Our team will call you within two hours”). Then someone has to keep that promise.

What goes into the knowledge base

Here’s a starter list of documents and entries to write before launch. Plain, short, current:

  • Rate card in sentences, with extras explained

  • Serviceability and transit times by zone

  • Pickup rules: cut-off times, booking, what to prepare

  • Tracking guide and a status glossary

  • Redelivery, address change and delivery-attempt policy

  • Prohibited and restricted items, and packaging guide

  • Claims process, deadlines and documents

  • Branch addresses, timings and contact numbers

  • Business and franchise enquiry information

  • A “hand-off rules” entry listing what the bot must never decide

Glanceia answers from content you upload, such as PDFs, Word documents, CSVs and FAQs, so every answer stays specific to your service. Keep one version of each document, delete old rate cards and date everything. A bot quoting last year’s rates is worse than no bot.

Courier-specific safety rules

Couriers handle names, addresses and phone numbers all day, so a chatbot adds a few responsibilities of its own.

Never ask for or accept delivery OTPs, passwords or payment details in chat. Say so in the bot’s rules and tell customers plainly: “We never ask for OTPs or card details here.”

Collect only what you need. A name and phone number is usually enough for a callback. Don’t ask for ID numbers.

Be clear that it’s an AI assistant, and that only your team can approve refunds, compensation or changes.

Mind India’s data protection law. The DPDP Act’s main duties apply from mid-2027, and a chatbot that collects names and phone numbers is within its scope. We cover it in Is Your Website Chatbot Compliant With India’s DPDP Act?.

Don’t let the bot be talked into promises. Customers can trick AI assistants into offering things they shouldn’t. Keep it informational and read Prompt Injection and Business Chatbots for how.

What the big names did

A few examples, all published by technology vendors, so treat them as indicative rather than independent proof.

Aramex moved part of its delivery support to WhatsApp self-service, handling about 500,000 messages a month and reporting 2.7 times the support productivity of a phone call, according to its messaging vendor’s customer story. It had set a goal of cutting calls by up to 80%, but the story reports productivity, not a final reduction.

Blue Dart used a WhatsApp chatbot aimed at cutting the number of times customers weren’t present when the delivery partner arrived, per a case study from conversational-AI company Haptik that reports a 2.5x return on investment. We haven’t verified how that figure was measured.

The shared lesson is the approach: put routine answers in front of customers first, and keep people for the exceptions. One more point from the same glossary of terms: silence is what triggers many status contacts, so clear proactive updates (SMS or WhatsApp at key milestones) reduce calls before they start. That’s a separate system from a website chatbot, and the two work well together.

Setting it up with Glanceia

With Glanceia, setup has three steps: connect your data, customise the look and tone, and add one line of code to your website. It works with platforms such as WordPress, Shopify, Wix and Squarespace.

  • Write a first message for each page. On the tracking page: “Not sure what a status means? Ask me.” On the rates page: “Want an estimate? Tell me the weight and pincode.” On the pickup page: “Questions before booking?” Glanceia lets you set a custom first message.

  • Capture details for handoffs. Glanceia can collect names, emails and phone numbers during chat and route lead data and important conversations to your inbox.

  • Use alerts for bookings. It can send real-time alerts when users reach key pages such as booking or payment.

  • Reach customers in their language. It supports over 95 languages.

  • Review the analytics for top questions and drop-off points.

What the site doesn’t describe: live shipment-status lookup by consignment number, specific fields such as a consignment number in the lead form, or a WhatsApp channel. If any of those matter to you, ask the team through the contact page before you promise them to customers. The site lists a Free plan with limits and a Pro plan marked “Coming Soon,” so check the pricing page for current details.

A 14-day pilot

You don’t need a big launch. Run a small, measured trial.

Days 1 to 2: Tally your calls and note your current weekly call count and the biggest categories.
Days 3 to 5: Write answers for the top five categories, plus handoff rules. Upload your rate card, glossary, pickup and claims documents.
Day 6: Test it as a stranger. Ask each of the ten questions, with typos and angry phrasing, and try to trick it into offering a refund.
Day 7: Go live on two pages only, such as tracking and rates, with a clear “talk to a person” route.
Days 8 to 13: Read the chats daily. Add missing answers and fix confusing ones.
Day 14: Compare your numbers with day one, then decide what to expand.

What to measure

  • Calls per week before and after

  • Contacts per 100 shipments

  • WISMO share, with a consistent definition

  • Questions the bot couldn’t answer

  • Time to first human reply on handoffs

  • Claims and missed-delivery contacts, which show whether you’re fixing causes

One trap: if other contact types grow faster, your WISMO percentage can fall even as the number of status calls rises. Watch both the share and the count.

Mistakes to avoid

  • Launching without a status glossary. It’s the cheapest call-saver there is.

  • Letting the bot guess about a specific parcel. “I can’t see your shipment” is a good answer.

  • Promising delivery dates.

  • Collecting OTPs or payment details.

  • Hiding the human option.

  • Outdated rates. Update the bot whenever prices or cut-offs change, especially before festival and sale periods.

  • No owner for handoffs. Someone must check the inbox and keep the promise.

  • Measuring only fewer calls. If customers leave unhappy, it isn’t a saving.

FAQ

Can a chatbot reduce support calls for a courier company?
It can take over repeated, factual questions: rates, transit times, pickup rules, tracking help and claims steps. How much it reduces calls depends on your mix of contacts and the quality of the content behind it.

Can it tell customers where their parcel is?
Only if it’s connected to your tracking system. Otherwise it can explain how to track and what statuses mean, then pass the customer on. Glanceia’s site doesn’t describe live tracking lookup, so ask the team if you need it.

What should it never do?
Guess about a specific shipment, promise delivery dates or refunds, ask for OTPs or payment details, or handle distressed customers alone.

Can it bring in business?
Yes. Glanceia can capture names, emails and phone numbers in chat and send lead data to your inbox, which suits seller and bulk-shipping enquiries.

Does it work in Indian languages?
Glanceia supports over 95 languages. Test the ones your customers use.

Do I need a developer?
No. You upload your content, customise the look and first message, and add one line of code to your site.

How much does it cost to start?
Glanceia has a free plan with no credit card required. See the pricing page for current details.

How do I know it’s working?
Compare calls per week and contacts per 100 shipments before and after, and read the chat history. Glanceia’s analytics show top questions and drop-off points.

Give your hub phones a break

You don’t have to automate everything. Start with the five easiest calls, write clear answers, keep a human for the rest, and measure for two weeks.

Glanceia is free to start with no credit card required. See the docs to get set up, or contact the team to talk through your delivery operation.

Found it useful?Share it with a friend.
Share this article

Related articles