Ecommerce Chatbot: The Complete Guide for 2026

An ecommerce chatbot answers your shoppers’ questions on your store, and when it’s connected to your store, it handles small jobs for them: finds a product, looks up an order, cancels one that hasn’t shipped. A good one answers from your catalog and policies, shows where the answer came from, and hands anything unusual to a person.

This guide covers the whole topic in one place: what an ecommerce chatbot should handle, how it works, what it costs, how to choose one, and how to set it up. We make ZynfoAI, so our product appears in the examples. Where we compare others, the facts come from their own pricing pages.

What an ecommerce chatbot should handle

Open your support inbox and you’ll see the same handful of questions. These are the ones a chatbot should take off your plate.

Order status. “Where is my order #1042?” is the question you’ll get most. The chatbot should look the order up in your store and show the status in the chat, not send a link to a tracking page.

Product and sizing questions. “Does the linen shirt run small? I’m usually a medium.” The answer should come from your size guide and catalog, with the source shown.

Finding the right product. “Something light for a hot sleeper, king size.” The chatbot should match a vague need to your catalog and show a product card the shopper can add to the cart.

Returns and cancellations. “Can I return the boots if they don’t fit?” gets your policy. “I ordered the wrong size an hour ago” gets the order cancelled, if it hasn’t shipped and the shopper confirms.

Out-of-stock items. “Will the trail runners be back in a 9?” The chatbot can’t promise a date, but it can save a back-in-stock request.

Anything that needs judgement. A chipped vase, a refund outside policy, a wholesale enquiry. The chatbot’s job here is a clean handoff to your team.

We’ve written out each of these as a full conversation in AI chatbot for ecommerce: 8 examples.

How an ecommerce chatbot works

There are three kinds, and the difference matters when you buy.

	Answers from	Reads your store	Takes actions
Rule-based bot	A fixed script	Sometimes	Rarely
AI chatbot	Your pages and files	No	No
AI agent	Your pages, files and catalog	Yes	Yes, the ones you allow

Most stores now want the third. When a shopper asks a question, the agent searches your sources (not the open web), writes the answer from what it found, and shows the source. If nothing covers the question, it sends your fallback message or offers a person. For order questions it asks for the order number and reads the order from your store.

For the plain-English version, read what is an ecommerce chatbot. For what an agent should and shouldn’t be allowed to do, see AI agents for ecommerce.

What an ecommerce chatbot costs

Pricing models differ more than features. These are the common ones, with list prices from each vendor’s own page as of October 2026:

Flat plan. ZynfoAI is $99 a month or $990 a year, with 5,000 conversations, 11 seats and the Shopify app included (pricing).
Free, with limits. Shopify Inbox is free and describes itself as “an AI-powered sales associate on your storefront”, for Shopify stores only.
Per conversation. Tidio lists Starter at $24.17 a month on annual billing for 100 conversations, with its Lyro AI agent from $32.50 a month.
Per AI resolution, plus seats or tickets. Gorgias charges $1.00 per AI Agent interaction on monthly plans. Intercom’s Fin is $0.99 per outcome, with a 50-outcome monthly minimum.

Price your busiest month, not an average one. Per-resolution pricing looks cheap at low volume and climbs fast in peak season. Our pricing breakdown walks through the maths, and the best ecommerce chatbots in 2026 compares six tools side by side.

How to choose one

Ask every vendor the same questions:

Does it answer only from your content, and show the source?
Can it look up a real order in your store platform?
Which actions can it take, and which can’t it?
What happens when a person is needed, and outside your working hours?
Which channels does it cover: website, WhatsApp, email, social?
Does it reply in your customer’s language?
What will it cost in your busiest month?
Can you see what it couldn’t answer?

Our buyer’s guide to the AI chatbot for ecommerce explains each question and how ZynfoAI answers it, including what it doesn’t do.

How to set up an ecommerce chatbot

Setup is mostly content and rules, not code. These steps follow ZynfoAI, but they apply to most tools.

1. Collect the answers shoppers already ask for

List the last 50 customer questions from your inbox. For each one, find where the answer lives. If it isn’t written down, write it now. Most stores find gaps: a size guide that lives in someone’s head, a returns policy that doesn’t mention sale items.

2. Add your sources

Add your website and the agent reads your pages. Upload what isn’t on the site: PDF care guides, Word documents, spreadsheets of product specs. Connect Google Drive if your team keeps docs there. Leave out old promotions and outdated policies. They confuse the agent the way they’d confuse a new hire.

3. Connect your store

Install the ZynfoAI Shopify app. Your catalog syncs into the agent’s knowledge, and the agent can recommend products and look up orders. Resync is manual: click Resync after a new collection or a price change.

4. Write the rules

In the AI Agent settings, brief the agent like a new team member:

“Ask for the order number before looking anything up.”
“Always confirm with the shopper before cancelling an order.”
“Never promise a delivery date. Quote the shipping policy instead.”

Then set a fallback message for questions your sources don’t cover. That’s what stops the agent guessing.

5. Choose the actions

The agent can cancel unfulfilled orders, log refund and return requests, share discount codes and save back-in-stock requests. Start with lookups and add the rest once you’ve read a week of conversations. Cancellation should always wait for the shopper’s yes. It only applies to orders that haven’t shipped, which matches how Shopify handles cancellations.

6. Set up the handoff

Turn on human handoff, add your working hours and choose round-robin, least busy or broadcast routing. When a shopper needs a person, the agent writes a summary and routes the chat to a teammate in the shared Inbox, and stops replying while they handle it. Unclaimed chats return to the queue. Outside hours, the shopper can leave a message, open a ticket or keep talking to the AI.

If shoppers message you on WhatsApp, connect your number through Meta’s Embedded Signup. WhatsApp and website chats land in the same Inbox. More in our WhatsApp AI chatbot guide.

7. Test like a difficult shopper, then go live

In the Playground, ask a vague question, try to cancel a shipped order, ask about a return outside the window, ask for something you don’t sell, and ask in another language your customers use. When the answers hold up, add the chat widget with one script tag or turn it on from the Shopify app.

After launch, read Unanswered Questions and Trending Topics every week. Fix the page, resync, and the next shopper gets the answer.

Guides by platform and store type

Shopify: how to add an AI chatbot to Shopify, and what the Shopify integration connects.
Shopify stores: the product walkthrough on ZynfoAI for online stores.
Shops with branches: the retail chatbot page covers store hours, returns and stock requests.
Webshops selling across borders: a chatbot for webshops that answers in the shopper’s language.
Restaurants: what an AI food ordering chatbot can and can’t do.
Notion users: train an AI agent on Shopify, Notion and WhatsApp.

What an ecommerce chatbot won’t fix

Gaps in your content. If the size guide is missing, so is the answer.
Checkout problems. Baymard Institute puts the average documented cart abandonment rate at 70.22%, across 50 studies. A chatbot can answer the question that stopped a shopper. It can’t fix surprise shipping costs or a slow checkout.
Your search rankings, directly. Google now shows FAQ rich results only for well-known government and health sites. The useful link is indirect: the questions your chatbot couldn’t answer tell you what to add to your product and policy pages.
Exceptions. Refunds outside policy, damaged items and complaints still belong with your team.

Common mistakes

Launching with only the homepage as a source.
No fallback message, so the agent has nothing safe to say when it doesn’t know.
Handoff turned on with no working hours set.
Forgetting to resync after a sale or a new collection.

Frequently asked questions

Which ecommerce chatbots handle order tracking and after-sales support?

Look for one that reads order data from your store rather than linking to a tracking page. ZynfoAI does this for Shopify, and also cancels unshipped orders, checks refund status and logs return requests. Gorgias and Tidio also connect to Shopify. Our comparison covers the differences.

Can I set one up without code?

Yes. Install the Shopify app or paste one script tag, and add your pages and files. Rules are written in plain language.

What’s an affordable chatbot for a small online store with order tracking?

ZynfoAI is $99 a month with order lookup, 5,000 conversations and 11 seats included, and a 7-day free trial. If you’re on Shopify and only need pre-sale answers, Shopify Inbox is free.

Can a chatbot handle a large product catalog?

Yes, as long as the catalog is synced and the product data is good. The chatbot can only recommend what your product titles and descriptions describe, so thin descriptions mean thin answers.

Will it replace my support team?

No. It takes the repeat questions so your team can spend time on the ones that need judgement.

For the eight buying questions in detail and how ZynfoAI answers them, see the AI chatbot for ecommerce buyer’s guide.