Artificial intelligence (AI) is becoming increasingly popular for assisting in shopping decisions, whether it’s finding the best deals on sneakers or comparing product specifications. Retailers are now exploring the possibility of AI agents making purchases on behalf of consumers without manual confirmation.
For instance, Amazon’s Alexa for Shopping in the U.S. now offers a feature where products can be automatically purchased when they reach a specified price threshold. Similarly, Robinhood recently introduced a new capability allowing AI agents to execute stock trades for users. Google has also unveiled a feature in its payments protocol that enables AI-driven tools to make purchases based on predefined criteria, with the functionality set to debut in the U.S. through Google’s new AI agent named Gemini Spark.
To ensure responsible purchasing, Google emphasizes the implementation of strict controls when allowing AI to buy products, likening the process to giving a teenager a debit card with limitations, as mentioned by Josh Woodward, vice-president of Google Labs. While these advancements represent initial steps in automated shopping, cybersecurity and industry experts highlight challenges related to privacy, standards, and consumer trust that must be addressed.
Despite the progress, concerns remain regarding the complexity of agentic checkout processes, such as navigating various payment options, validations, and security measures. The need for consumer consent is also paramount to avoid potential errors by AI agents in making purchasing decisions. Cybersecurity experts stress the importance of safeguards to prevent unauthorized purchases from unsecured websites and potential exploitation by malicious entities.
As the industry prepares for widespread adoption of agentic payments, initiatives like Visa’s Agentic Ready program in Canada aim to test and refine these systems collaboratively with financial institutions. However, consumer acceptance of AI agents handling transactions may take time, with research showing mixed comfort levels among consumers regarding autonomous purchases facilitated by AI.
While the convenience of AI-driven shopping is evident, experts suggest that certain product categories, particularly low-cost and replenishable items, are more likely to be outsourced to AI agents. Yet, the enjoyment and personal expression associated with traditional shopping experiences may hinder widespread adoption of agentic shopping across all product categories. Ultimately, the balance between automation and the human touch in shopping experiences remains a critical consideration for the future of retail.