Add an AI Shop Assistant to WooCommerce with ShopChat

It’s 11pm and a shopper wants to know whether the jacket runs large and if it ships to Canada. Nobody is online, the contact form promises a reply “within 24 hours”, and the tab gets closed. Multiply that by every product page in your catalogue and you have a quiet, steady leak.

It doesn’t have to work that way.

ShopChat — AI Shop Assistant for WooCommerce adds a chat widget that reads your live product catalogue, answers shopper questions, suggests relevant products and hands the conversation to a human when it should.

This guide walks through a complete WooCommerce AI assistant setup, from install to your first test questions, so you can have it running in about 10 minutes.

Why a shop assistant beats a contact form

Most pre-sale questions are small and repetitive: sizing, stock, delivery times, compatibility, returns. Each one is easy to answer, but only if someone is there at the right moment. A shopper with a question and no answer rarely waits. They open another store.

An assistant that knows your catalogue answers on the spot, at any hour, and keeps your team free for the conversations that need a person. If you want the wider picture first, our guide to adding an AI chatbot to WooCommerce covers the options. Here we focus on the hands-on setup.

How to set up ShopChat in 7 steps

1. Install and activate the plugin
   Upload the ShopChat zip under Plugins → Add New → Upload Plugin, activate it, and enter your licence key when prompted.
2. Connect your AI provider
   Create an API key with your AI provider, for example OpenAI or Anthropic, and paste it into the ShopChat settings. You pay the provider directly for usage, so set a monthly spending limit on their side.
3. Let it read your catalogue
   ShopChat works from your live WooCommerce products, so prices, stock status and descriptions stay current without a manual FAQ import. Open a few products you know well and make sure their descriptions are accurate, because the assistant is only as good as the content it reads.
4. Set the tone and the limits
   Describe how the assistant should sound (friendly, brief, on-brand) and what it must not do: promise discounts you don’t offer, give legal or medical advice, or guess at delivery dates. Clear limits prevent most awkward answers.
5. Add your upsell rules
   Decide when the assistant may recommend a related product, such as an accessory, a bundle or a higher-spec alternative. Keep it relevant and occasional. A suggestion that answers the shopper’s question feels like service; one on every message feels like spam.
6. Configure the human handoff
   Choose when the conversation passes to your team: when the assistant isn’t sure, when the shopper asks for a person, or for topics like order problems. Tell shoppers your support hours so the handoff feels honest.
7. Test with real questions
   Pull 10 questions from your support inbox and ask them in the widget. Include a sizing question, a stock question, a shipping question, a returns question and one that is completely off-topic. Adjust the tone and limits until the answers sound like your store.

What to configure and why

Keep improving it after launch

The first week is where you learn the most. Read the conversations, note the questions the assistant struggled with, and fix the cause. Usually that means a product description that is vague, a shipping policy that is missing, or a limit that needs to be clearer.

• Add missing sizes, materials and compatibility details to product pages.
• Publish a clear shipping and returns page the assistant can draw on.
• Review handoff conversations weekly to spot gaps.

For the bigger trends behind this, see AI for WooCommerce stores in 2026.

Conclusion

A shop assistant works when it knows your products, speaks like your brand and knows when to step aside. Seven short steps get you there, and a handful of real test questions tell you whether it is ready.

Set it up once, review the first week of conversations, and let it answer the 11pm questions your team can’t.