Deciding whether or not you can trust a website's photo, video, and/or description of its products is one of ecommerce's biggest issues.
A video can make a pair of shoes look great, a creator can recommend a skincare brand, and an algorithm can put a product in front of exactly the right person at exactly the right time.
But the questions that stop a shopper from checking out tend to come later. Will it fit? Is it right for their needs? Is there a better alternative? Can they trust the recommendation?
For Zhengyang Yang, a Product Designer with TikTok, that is where shopping AI has a useful role to play.
“The job that AI plays is to help users from inspiration to confidence,” he told CX Today.
That is an important distinction. What retailers do not need is another digital destination for customers to learn, navigate, and eventually abandon. Instead, they need assistance that helps them make a decision without taking the decision away from them.
Yang spoke to CX Today in a personal capacity. His views do not represent official TikTok product strategy or policy.
AI Should Remove Doubt, Not Add More Clicks
Digital shopping journeys are already filled with search bars, filters, product pages, ratings, reviews, creator videos, recommendations, promotions, and ads competing for attention.
Dropping a chatbot or AI summary into every stage of that journey may look innovative, but it can easily become one more thing for the customer to process.
“The role of AI should reduce cognitive load instead of just providing a lot of more clicks,” Yang said.
That means shopping AI needs a clear job.
A customer considering a shirt, for example, may understand the product but still be unsure about sizing. They could tell an assistant which size they wear in other brands and ask for a recommendation.
That is a useful intervention because it addresses a specific uncertainty that a video or product description may not answer.
The same applies when customers are comparing similar products, trying to interpret a large volume of reviews, or weighing up complex specifications.
The key is not to add AI because a product team can; it is to put it where it removes friction from a decision the customer is already trying to make.
Do Not Make Customers Learn How to Shop Again
The other potential risk at play here is that in the rush to make shopping more conversational, brands may push customers out of familiar flows and into AI experiences they never asked for.
Most online shoppers already know their standard ecommerce routine. Yang believes AI should work within that established behavior rather than replacing it with a wholly new journey:
“If users want to have a comparison in the shopping cart, you can add AI to the shopping cart. And when the users may have some questions or when they want to see some frequently asked Q&As on the product detail page, you can also add AI to that position.”
This “progressive assistance” model has a clear appeal. The technology is available when it can help, but it does not demand that the customer start from scratch in an unfamiliar chat interface.
As Yang put it:
