A shopper asks an AI assistant whether your store delivers a particular product to their area. The answer may shape their decision before they visit a page. Shopify’s Knowledge Base app, launched in May, lets merchants view and customise FAQs used by AI shopping agents. It gives the store team a practical place to review its answers and notice the questions people are bringing to these conversations.

Choose facts somebody can stand behind

Start with the questions support already recognises: delivery destinations, dispatch expectations, product compatibility and what happens when an item is unsuitable. Name the person responsible for each answer. The catalogue owner may know the product details, while the fulfilment team owns the delivery promise.

Avoid turning an uncertain answer into a cheerful universal claim. If a service depends on the destination or item, describe the dependency and the useful next step. Good content design helps a shopper understand the boundary of an answer. It also gives the team something concrete to maintain when the business changes.

Keep the source information in agreement

Review the storefront, product data and written policies alongside the FAQs. If the product page and the support answer disagree, adding another channel gives that contradiction another place to travel. Resolve the underlying fact before polishing the wording for an assistant.

For example, a compatibility answer should identify the relevant model or measurement and the evidence behind it. A photograph may suggest a fit without establishing one. Ask whether a customer can verify the answer on the store’s own page, and make the route to that information clear when a longer explanation is needed.

Make a useful answer traceable: Recognise the shopper’s question; Check the business-owned source; Keep page and FAQ in agreement; Review after a change
Connect a shopper’s question to an owned business fact, consistent storefront information and a review routine.

Use the questions to improve the website too

The app’s view of questions requested by AI agents can provide ideas for better explanations. Treat these as clues to investigate with the people serving customers. A repeated question about delivery may point to an unclear product page, an incomplete policy or a genuinely unusual buying situation.

Choose one question and improve its full journey. Refine the FAQ, update the relevant page and check the support response. This makes the work useful even when the customer arrives through ordinary search, a shared link or a conversation with the team. One maintained answer can support several ways of finding the business.

Review what changed before widening the effort

Begin with a small collection of answers and a review date. When an offer, range or service changes, include those answers in the update routine. A business-owned knowledge base is an ongoing content responsibility, not a file to approve once and forget.

An accurate FAQ cannot guarantee that every AI service will recommend a store or repeat an answer correctly. The value under the merchant’s control is clearer information, a better route to evidence and a team that knows who maintains it. Start there, and use real questions to decide what deserves attention next.