AI for ecommerce

What is AI used for in an online store?

Stores use it mainly to serve customers without growing the team and to keep a large catalog up to date. In a store that ships 50 to 1,000 orders a day, support fills up with questions about shipping status, size exchanges and returns. All of them have an answer in your platform or on the carrier's site, but someone has to look it up.

The catalog works the same way: uploading 300 new products with descriptions, attributes and SEO text takes weeks. AI produces the first draft in hours and your team reviews it.

Which ecommerce automations have the most impact?

These have the biggest effect in small and mid-sized online stores:

• Order and shipping status: the agent identifies the customer, checks the order in the store and the carrier's tracking, and replies by chat, email or WhatsApp. See AI customer service.
• Returns and exchanges: checks whether the order is inside the return window, explains the process, generates the label if your system allows it and records the reason.
• Product listings at scale: descriptions, titles, bullets and metadata from the supplier sheet and photos, in your tone and in several languages.
• Catalog enrichment: fills in missing attributes (material, dimensions, compatibility) so filters and search work.
• Recommendations: the agent suggests products based on what the customer asks, using live stock.
• Personalised abandoned cart messages: emails or WhatsApp messages that answer the likely doubt (shipping, size, delivery time) instead of a generic reminder. See AI email automation.

Does it work with Shopify, WooCommerce and BigCommerce?

Yes. All three platforms have APIs to read orders, customers and products, and that is what the agent uses. Shopify is the easiest to integrate; WooCommerce depends more on the plugins and version you have installed, and BigCommerce sits in between.

We also connect carriers (through their API or your shipping aggregator), helpdesk tools such as Gorgias or Zendesk, and email platforms such as Klaviyo or Mailchimp. If your store is custom-built, we work with its database or with the API it exposes. When you need something more, we cover it with custom software development.

What should AI not decide in an online store?

AI does not approve refunds outside policy and does not invent promotions. The agent applies your returns policy exactly as written, including any statutory rights such as the 14-day cancellation period for distance sales in the UK and EU, and it passes every exception to a person. Rules differ by country and state, so we set the policy logic with you.

In product listings, AI must not invent features. If the supplier sheet does not say a fabric is waterproof, the listing will not say it either. That is why we review a sample from each batch before publishing and flag products with incomplete information instead of filling gaps.

A Monday of support in a fashion store

A fashion store on Shopify with two support agents starts Monday with 140 messages from the weekend. The assistant has resolved 95 with no intervention: 60 order status queries, 25 returns within the window with their labels and 10 sizing questions answered from the size guide.

The remaining 45 sit in the team's inbox, sorted by priority: lost parcels, damaged goods and a customer asking for an exchange after the window closed. Each ticket arrives with the order, the history and a draft reply.

In the afternoon, the merchandising team uploads 120 items from the new collection. AI generates the listings in English and French; the manager reviews 20 at random, corrects two and publishes the batch. The scenario is illustrative.