Finding the right rain jacket
A customer is looking for a rain jacket for her daily commute. She wants to know whether it will keep her dry on a longer walk and which size will fit over a jumper. A search box can give her a list of jackets. Useful advice needs to understand which details matter for that journey.
That is where BigCommerce is starting. On 16 November, Commerce.com announced Shopping Agent for eligible merchants. It is designed to answer questions, compare products and guide customers towards the basket and checkout inside the store. The company says the rollout has begun for eligible merchants, without specifying all the regions and industries included.
What can BigCommerce Shopping Agent do?
According to Commerce.com, merchants can adapt the assistant to their brand. Customers can describe what they need in everyday language, ask follow up questions and compare products with relevant details. The agent is intended to add items to the basket during the conversation and provide checkout guidance. Those are the provider’s stated capabilities; their reliability still needs to be tested in a particular store.
Underneath is a new conversational storefront search. The company says it combines keyword relevance with semantic understanding and is available to all paid BigCommerce plans. That is a different availability statement from Shopping Agent itself, which the announcement limits to eligible merchants.
The merchant keeps the stage
Here the assistant lives on the merchant’s own storefront. That is a different starting point from a personal shopping agent comparing several merchants outside their stores. The customer is already with the merchant; the question is whether the experience helps her make a better decision there.
The merchant shapes the product information and presentation behind that advice. The team also needs to check that a recommendation reflects the right variant, current price and delivery terms. A persuasive answer about a rain jacket is not much help if the recommended size is out of stock.
Advice needs more than a chat
In Experience that improves a decision, I look at what a merchant can contribute to a choice. In the fashion demo, a jacket, jeans and cap form a coherent selection that leads to actual product details. Chat is one possible entry point. What matters is whether the advice leads to a decision the customer can understand.
A related example is J.Crew’s Style Advisor. It starts with an occasion and draws on looks assembled by the brand’s team. BigCommerce describes an adaptable capability for eligible merchants on its platform. Whether an individual merchant creates advice that is equally useful depends on its own products and content.
What about B2B buyers?
Commerce.com says it is also developing agents for business buyers. Its B2B Agent is in early access and is intended to support purchase order workflows. The company names complex catalogues, negotiated prices and account requirements as further areas of work.
A business buyer might need a different answer altogether: does this spare part fit the machine revision, does the agreed price apply, and is this person allowed to order that quantity? Those are business process questions. The announcement does not establish that every one of those steps is already available in the consumer Shopping Agent.
How should a merchant test the shopping agent?
I would start with ten real questions from sales and support. For each answer, check whether the recommendation fits, the variant is right and the customer can see price, availability and delivery terms. Then try a question where the necessary detail is missing. Asking for clarification can be better advice than a confident guess.
Next, follow the complete path into the basket. My guide to GEO and AI search explains how teams can check product information against real customer questions. Its test is aimed at external search; the same care with measurements, variants and sources helps when assessing an assistant inside the store.
Source and further reading
Read the original announcement ↗
