Wizard and Mastercard Launch In‑Chat AI Shopping That Completes Purchases

TL;DR: Wizard Commerce and Mastercard announced a joint AI shopping tool that lets users complete purchases inside the chat, addressing the fact that about 90% of current AI shopping assistants only provide conversation without checkout. The system combines Mastercard’s Insight Tokens for personalized recommendations and Agent Pay for seamless, in‑chat payments.

Most AI shopping assistants today act like a knowledgeable store clerk who can’t ring up a sale. They can describe products, compare prices and even suggest trends, but they stop short of letting the user buy anything. This leaves shoppers clicking out to external sites, where many abandon the process.

Wizard CEO Melissa Bridgeford cited the "91% problem" in a recent interview, noting that nine out of ten AI shopping tools deliver a chat experience without a shoppable one. The gap creates friction similar to walking into a store where a helpful associate can explain a product but there is no cash register to complete the transaction.

The partnership tackles this with three components. First, Mastercard’s Insight Tokens use aggregated spending data to tailor recommendations, working like a salesperson who sizes up a shopper the moment they enter. Second, Wizard’s own search engine refines product matches using relevance, quality and brand signals. Third, Mastercard’s Agent Pay enables the AI to initiate and finish a purchase inside the chat using tokenized credentials, removing redirects and extra steps.

By keeping the entire journey inside the conversation, the tool aims to boost conversion rates that currently drop when users must leave the chat. Bridgeford argues that the system ranks products based on internal scoring rather than paid placement, hoping to alleviate consumer skepticism about biased suggestions. Merchants could see higher completion rates if shoppers trust the AI’s curation.

Watch for adoption metrics in the coming quarters, how often users complete purchases without leaving the chat and whether trust scores improve as the technology scales.