Table of Contents
- Key takeaways
- What is an ecommerce chatbot?
- What does an ecommerce chatbot do?
- Why do ecommerce stores use chatbots?
- How to set up an ecommerce chatbot in 8 steps
- 1. List the questions shoppers actually ask
- 2. Fix the content the bot will learn from
- 3. Decide which jobs need store data
- 4. Pick the tool by platform and order needs
- 5. Train the bot and set its rules
- 6. Test it on 50 real questions before launch
- 7. Install the widget and set the handoff
- 8. Measure four numbers every week
- The stack: which tool does which job
- 1. Order layer: Shopify Inbox or Gorgias
- 2. Knowledge-trained answer layer: SiteGPT
- 3. Human layer: your team
- How much does an ecommerce chatbot cost?
- Common ecommerce chatbot mistakes
- Frequently asked questions
- Basics
- What is an ecommerce chatbot?
- What is the difference between an ecommerce chatbot and live chat?
- Setup
- How long does it take to set up an ecommerce chatbot?
- What content should an ecommerce chatbot be trained on?
- Orders and integrations
- Can an ecommerce chatbot track orders?
- Does SiteGPT integrate with Shopify?
- Cost
- Is there a free ecommerce chatbot?
- How much does SiteGPT cost for a store?
- Sources
Created by
Do not index
Created time
Sep 9, 2023 12:17 AM
An ecommerce chatbot is software that answers shoppers in a chat window on an online store: it handles product, sizing, shipping and returns questions from your content, and it can track or change orders only when it is connected to the store platform or a helpdesk.
TL;DR
- An ecommerce chatbot does two kinds of work. Answer jobs (product, sizing, shipping, returns) need only your content. Order jobs (order status, address changes, refunds) need a connection to the store.
- Set one up in eight steps: list the top questions, fix the content, pick the tool, train it, test it on 50 real questions, install it, set the handoff, then measure weekly.
- Pick the tool by two facts: which platform runs the store, and whether the bot must act on orders. The ranked list of ecommerce chatbot solutions compares five options.
- Plan the handoff from day one. In a 2026 Gartner survey, only 27% of customers said they would try a chatbot again after a bad experience.
- An answer chatbot such as SiteGPT costs $59 a month billed monthly, or $39 a month billed yearly. Shopify Inbox is free on Shopify stores.
Key takeaways
Question | Short answer |
What does an ecommerce chatbot do? | Answers product, sizing, shipping and returns questions, and gives order updates when connected to the store |
Which jobs need store data? | Order status, order changes and refunds. Everything else can come from your pages and policies |
How long is the setup? | Eight steps. The content and testing steps take the most time |
What should it cost to start? | $0 for Shopify Inbox on Shopify, or $59 a month billed monthly for SiteGPT Starter on any platform |
What decides success? | A narrow, tested scope and a visible route to a person |
What is an ecommerce chatbot?
An ecommerce chatbot is a chat assistant on an online store that answers shoppers before and after they buy. Most current ecommerce chatbots use AI to write each reply from a source of truth, such as product pages, store policies or order data.
The source decides what the bot can do. Two types cover most stores:
- Answer chatbot. The bot is trained on your content: product pages, shipping and returns policies, size guides, manuals and FAQs. It installs with a script on any website. It cannot see orders.
- Store-connected chatbot. The bot reads the store itself, such as the catalog, inventory and orders. Shopify Inbox works this way on Shopify, and the Gorgias AI Agent works inside the Gorgias helpdesk.
Criterion | Answer chatbot | Store-connected chatbot |
Learns from | Pages, policies, files, drives, videos | Catalog, inventory, policies, orders |
Installs on | Any website, by script | Only the platforms it integrates with |
Order status | Links to the tracking page or hands off | Can give order updates |
Example | Shopify Inbox, Gorgias AI Agent |
For store names and the results they published, see ecommerce chatbot examples. For uses outside retail, see chatbot use cases.
What does an ecommerce chatbot do?
An ecommerce chatbot does seven common jobs. Five of them need only your content. Two need a connection to the store.
Job | Example shopper question | Where the answer comes from | Needs store data? |
Product questions | "Is this jacket waterproof?" | Product pages, manuals, spec sheets | No |
Sizing and fit | "Does this run small?" | Size guides, fit notes | No |
Shipping | "How long does delivery to Canada take?" | Shipping policy | No |
Returns and exchanges | "Can I return a sale item?" | Returns policy | No |
Product finder | "Which blender is best for smoothies?" | Product pages, or the live catalog | Only for live stock and prices |
Order status | "Where is my order?" | Order and tracking data | Yes |
Order changes and refunds | "Can I change my delivery address?" | Order data with edit rights | Yes, or a person |
A chatbot also hands the conversation to a person when it cannot help. Damaged items, payment disputes and complaints belong with a person from the start.
Why do ecommerce stores use chatbots?
Ecommerce stores use chatbots to answer pre-sale questions at the moment a shopper decides, and to take repeat questions off the team. The pre-sale moment matters because most carts never reach checkout. Baymard Institute puts the average documented cart abandonment rate at 70.22%, across 50 studies.
The risk runs the other way too. Gartner surveyed 3,566 customers in February and March 2026. 49% said they would use a company chatbot if offered one, but only 7% used one in their most recent service contact. Only 27% would try a chatbot again after a bad experience.
Gartner's advice was to "prioritize reliability over reach". For a store, that means launching with the questions the bot can answer correctly every time.
How to set up an ecommerce chatbot in 8 steps
To set up an ecommerce chatbot, work through these eight steps in order. Steps 1, 2 and 6 decide most of the result.
1. List the questions shoppers actually ask
List the questions shoppers actually ask by exporting the last 90 days of chat, email and contact-form messages. Group them into product, sizing, shipping, returns, order status, payment and other. Count each group.
The count shows the scope. If order status makes up a large share of the messages, the bot needs store access or a clear link to the tracking page.
2. Fix the content the bot will learn from
Fix the content before you train anything, because the bot can only repeat what your pages say. Check five sources:
- Shipping policy: carriers, delivery times by region, costs and cut-off times
- Returns policy: window, condition rules, sale-item rules and how to start a return
- Product pages: materials, dimensions, compatibility and care
- Size guides and fit notes for every product line
- An FAQ page that answers the top questions from step 1 in plain sentences
Write the missing answers now. A question with no written answer becomes a guess or a handoff.
3. Decide which jobs need store data
Decide which jobs need store data using the table in the section above. If the bot must look up or change orders, you need a store-connected tool. If it only answers, an answer chatbot is enough, and it can link to the order tracking page for order status.
4. Pick the tool by platform and order needs
Pick the tool by two facts: the platform that runs the store, and the answer to step 3.
If you need | Pick | Why |
A free AI agent on a Shopify store | Shopify Inbox | Free, made by Shopify, connects to the catalog, inventory and policies |
A bot that works orders inside a ticket queue | Gorgias | Helpdesk built for ecommerce, with an AI Agent on every plan |
Answers from manuals, PDFs, drives or videos, on any platform | Trains on pages, files, Google Drive, Dropbox, OneDrive, Box, Notion, GitHub and YouTube | |
A store on WordPress, Wix, Squarespace or a custom build | Installs by embed script on any website |
For five ranked options with plans and prices, see the ecommerce chatbot solutions list. Shopify-only options are compared in the best AI chatbots for Shopify.
5. Train the bot and set its rules
Train the bot on the sources from step 2, then write its rules. Set four rules at minimum:
- The topics it answers, taken from step 1
- The tone, in one or two sentences
- What it says when the content has no answer: it declines and offers a person
- When it escalates: the shopper asks for a person, the issue is urgent, or the shopper is frustrated
Leave discounts and refund promises out of scope unless a person or a connected system approves them. Chatbot training covers source preparation in more depth.
6. Test it on 50 real questions before launch
Test the bot on 50 real questions from step 1 before any shopper sees it. Split them into three groups and score each group separately:
Group | Questions | Example result | Score |
Answerable from your content | 30 | 27 correct, with the right source page | 90% |
Not in your content | 10 | 8 declined, 2 guessed | 80% |
Should reach a person | 10 | 9 handed off | 90% |
Total | 50 | 44 handled correctly | 88% |
The numbers above are an example. In that example, the 2 guesses are the problem to fix first: each one is a wrong answer a shopper would see. Add the missing content or tighten the decline rule, then rerun the same 50 questions.
7. Install the widget and set the handoff
Install the widget on the pages where questions start: product pages, the cart, and the shipping and returns pages. Most tools install as an app or as a script pasted into the site template.
Then set the handoff. Decide who answers escalated chats, during which hours, and what shoppers see outside those hours. In SiteGPT, human support is available on every plan. It adds escalation buttons after each answer and lets you write your own escalation rules.
8. Measure four numbers every week
Measure four numbers every week, and read a sample of transcripts with them:
- Resolution rate: the share of chats that end without a person
- Escalation rate: the share handed to a person, and why
- Unanswered questions: the questions the bot declined, which show the content to write next
- Shopper rating: thumbs up or down on answers
Widen the scope only after these numbers hold for a few weeks. Chatbot best practices lists the checks behind each number.
The stack: which tool does which job
The stack for an ecommerce chatbot has three layers, and one tool rarely covers all three equally well. The table names each layer, its job, and the tool that fits it.
# | Layer | Job | Tool |
1 | Order layer | Order status, address changes, refunds | The store platform or a helpdesk, for example Shopify Inbox or Gorgias |
2 | Knowledge-trained answer layer | Product, sizing, shipping and returns answers from your content | |
3 | Human layer | Disputes, damaged items and anything the bot cannot answer | Your team, through the chatbot's handoff or a connected chat tool |
1. Order layer: Shopify Inbox or Gorgias
The order layer belongs to a tool that can see orders. Shopify Inbox is free and made by Shopify, and its app listing names order updates among its automated responses. It works on Shopify stores only. Gorgias is a helpdesk for ecommerce teams with an AI Agent on every plan. Its Starter helpdesk is $10 a month billed monthly for 50 tickets, and the AI Agent is billed when it resolves a conversation.
2. Knowledge-trained answer layer: SiteGPT
SiteGPT is the answer layer: an AI chatbot trained on your own content that installs by script on any website. It fits stores whose answers sit in product pages, manuals, size guides, PDFs and product videos, and stores that do not run on Shopify.
SiteGPT has real limits in this stack. Its integrations page lists no Shopify, WooCommerce, BigCommerce or Gorgias connection, so it does not look up or change orders. For "where is my order", it points shoppers to the tracking page or hands the chat to a person. There is no free plan, and the 7-day free trial requires a credit card.
3. Human layer: your team
The human layer is the person who replies when the bot cannot. SiteGPT's built-in human support works on every plan. From the Growth plan, the SiteGPT iOS app alerts the team when a shopper asks for a person, and chats can also go to Crisp, Slack, Messenger, Google Chat, Freshdesk Messaging or Zoho SalesIQ. Zendesk is on Scale and above.
How much does an ecommerce chatbot cost?
An ecommerce chatbot costs from $0 a month on Shopify to $59 a month billed monthly for an answer chatbot on any platform. The table pre-calculates the yearly cost.
Setup | Billed monthly | Billed yearly | What it includes |
Shopify Inbox | Free | Free | AI agent on one Shopify store |
$59 a month ($708 a year) | $39 a month ($468 a year) | 1 chatbot, 4,000 GPT-4.1-mini messages a month, 1,000 pages | |
$129 a month ($1,548 a year) | $79 a month ($948 a year) | 2 chatbots, 10,000 messages, chat integrations, iOS app, monthly auto refresh | |
Gorgias Starter helpdesk | $10 a month | Starter is billed monthly only | 50 tickets a month. AI Agent billed per resolved conversation |
Yearly billing saves $240 a year on SiteGPT Starter (34%) and $600 a year on Growth (39%). The SiteGPT quota is 4,000 GPT-4.1-mini messages on Starter, or 400 GPT-4.1 messages. For a full cost breakdown with setup and staff time, see how much a chatbot costs.
Common ecommerce chatbot mistakes
The most common ecommerce chatbot mistakes come from launching too wide or too early:
- Training on thin policies. A vague returns page produces vague answers. Fix the content in step 2 first.
- Promising what the bot cannot do. An answer chatbot cannot track orders, so the welcome message should not offer it.
- Hiding the person. 87% of customers in the 2026 Gartner survey called access to a human agent essential.
- Skipping the test set. Without step 6, the first wrong answers reach shoppers.
- Letting content go stale. New prices and policies must reach the bot the day they go live.
Try the setup on your own store with a 7-day SiteGPT trial.
Frequently asked questions
Basics
What is an ecommerce chatbot?
An ecommerce chatbot is a chat assistant on an online store that answers shopper questions about products, sizing, shipping and returns. When it is connected to the store platform or a helpdesk, it can also give order updates or act on orders.
What is the difference between an ecommerce chatbot and live chat?
An ecommerce chatbot answers automatically at any hour. Live chat connects the shopper to a person. Most stores run both, with the bot handing hard questions to a person. The full comparison is in chatbots vs live chat.
Setup
How long does it take to set up an ecommerce chatbot?
The install takes minutes, because most tools are an app or a pasted script. The content review in step 2 and the 50-question test in step 6 take the most time.
What content should an ecommerce chatbot be trained on?
An ecommerce chatbot should be trained on the shipping policy, the returns policy, product pages, size guides and an FAQ page. Manuals, spec sheets and product videos help for technical products.
Orders and integrations
Can an ecommerce chatbot track orders?
An ecommerce chatbot can track orders only if it is connected to the store platform or a helpdesk that holds order data. Shopify Inbox lists order updates on Shopify. An answer chatbot such as SiteGPT links shoppers to the tracking page or hands off to a person.
Does SiteGPT integrate with Shopify?
SiteGPT lists no Shopify integration on its integrations page. It installs on a Shopify store with an embed script and answers from your content, but it does not see orders.
Cost
Is there a free ecommerce chatbot?
Shopify Inbox is free on Shopify stores. SiteGPT has no free plan. It offers a 7-day free trial that requires a credit card.
How much does SiteGPT cost for a store?
SiteGPT Starter costs $59 a month billed monthly, or $39 a month billed yearly ($468 a year). Growth costs $129 a month billed monthly, or $79 a month billed yearly.
Sources
Sources checked on 18 October 2026.
Written by
