Facebook Messenger Chatbot for Online Stores: A Setup Checklist

For a lot of online stores, the Facebook Page inbox is still where most sales start. Someone sees a post, taps "Send message", and asks the price. If the reply comes two hours later, they have often already bought from the page that answered first.

A Messenger chatbot closes that gap. But the setup decides whether it helps or embarrasses you. This is the checklist we work through when we set one up for a store, in the order that saves the most rework.

1. Sort out access before anything else

Half of all setup delays come from access, not technology.

• Make sure you have full admin access to the Facebook Page, not just editor or moderator.
• If the Page sits inside a Meta Business account (Business Manager), check that you, or whoever is doing the setup, are an admin there too.
• Find out who originally created the Page. If that person left the company years ago, fix it now.
• Turn on two-factor authentication for every admin account. A hacked admin is far worse than a slow inbox.

Connection should always happen through Meta's official login, where you approve permissions on a Meta screen. If a tool asks for your Facebook password, walk away.

2. Write down what customers actually ask

Before you write a single answer, scroll back through the last few weeks of Messenger chats and list the questions. You will notice most of them repeat:

• How much is this?
• Is it available in my size or colour?
• Do you deliver to my city, and how much is delivery?
• How long will it take?
• Can I pay cash on delivery? Do you take eSewa, Khalti, card or bank transfer?
• Can I exchange it if it doesn't fit?
• Where is your shop?

This list is your starting scope. It also shows you the tone customers use. If half your messages arrive as "price kati ho?" or in Hindi or Arabic, your bot needs to handle that from day one, not as a later upgrade. We cover that in our guide to Romanized Nepali chatbots.

3. Build the knowledge the bot will answer from

An AI chatbot is only as accurate as the information behind it. For a store, that means:

• Products: name, price, variants (size, colour, storage), what's in stock, and a short description in plain words.
• Delivery: areas you cover, charges per area, typical delivery time, and what happens outside those areas.
• Payment: every method you accept and any conditions (for example, advance payment for custom orders).
• Policies: exchange, return and warranty rules, written the way you'd explain them on the phone.
• Shop basics: address, map link, opening hours, holidays.

Keep prices in one place that someone owns and updates. The most common reason a store bot gives wrong answers is not the AI, it's an old price list. Our chatbot knowledge base guide goes deeper on structuring this.

4. Decide the first message and the tone

The first reply sets expectations. Keep it short, say it's an automated assistant, and make it easy to reach a person.

Hi! I'm the shop's assistant. Ask me about prices, sizes or delivery, or type "agent" anytime to talk to our team.

Then match your brand. A streetwear page can be casual. A pharmacy or electronics store should be clearer and more careful. Avoid long welcome menus with eight buttons. Most people just want to type their question.

5. Set up order and lead capture

This is where the bot pays for itself. When a customer shows buying intent, the bot should collect what your team needs to confirm the order, in a natural back-and-forth rather than a form:

1. Item and variant
2. Quantity
3. Name
4. Phone number
5. Delivery address or area
6. Preferred payment method

The finished order should land somewhere your team actually looks, with the conversation attached. Decide now who confirms orders and how fast. A captured order that nobody calls back is worse than no bot at all. For more on this, see lead capture in chat.

6. Write the handoff rules

Decide which conversations go straight to a person:

• Complaints, damaged items and refunds
• Bulk or wholesale enquiries
• Anything the bot isn't confident about
• Anyone who asks for a human

Also decide what happens outside working hours. "Our team will reply after 10 am tomorrow" is fine. Silence is not. We explain this in detail in when your chatbot should hand off to a person.

7. Connect comments, not just the inbox

On Facebook, many buyers never message you. They comment "price?" under a post and wait. A good setup watches comments too: it can reply publicly and send a private message with the details. Meta has rules on private replies to comments, generally limiting how and when you can message a commenter, so check the current policy. The comment-to-DM guide explains how this works in practice.

8. Test like a difficult customer

Before going live, have three or four people on your team message the Page from personal accounts and try to break it:

• Ask about a product you don't sell.
• Ask in mixed language, with spelling mistakes.
• Ask two questions in one message.
• Ask for a discount.
• Complain angrily about a late delivery.
• Ask for a human.

Every wrong or awkward answer becomes a fix to the knowledge or the handoff rules. Expect a few rounds.

9. Watch the first week closely

Read conversations every day for the first week. Look for:

• Questions the bot couldn't answer (add them)
• Answers that were technically right but unhelpful (rewrite them)
• Handoffs that waited too long for a person (fix staffing, not the bot)
• Orders captured but not followed up

After the first couple of weeks, a weekly review is usually enough.

A quick scenario

Say you run a phone accessories shop in Butwal. Before the bot, your evening messages pile up and you answer them at midnight. After setup, someone asks at 9 pm whether a case fits their phone model and whether you deliver to Bhairahawa. The bot answers both, quotes the delivery charge, collects their address and phone number, and your staff confirm the order in the morning. Nothing magical, just no lost evening enquiries.

If you'd rather not do it alone

Orbit's team does this setup with you: we connect your Page through Meta's official API, build the product and policy knowledge, set handoff rules and test it before launch. You can see how the Facebook Messenger chatbot works, read about ecommerce setups, or book a demo and we'll walk through it using your own Page.