The introduction of agentic commerce platforms, exemplified by Mastercard’s Agent Pay, signifies a new stage in integrating artificial intelligence into the shopping experience. However, such systems are not yet widespread, and Mastercard is advancing plans for an Agent Pay pilot in the UAE.
Enhancing Personalized Shopping
These platforms intend to provide users with AI agents that function as personal shoppers. Users can delegate tasks from product selection through to transaction completion with minimal input.
Pilot Program Details
The Agent Pay pilot in the UAE will be conducted in collaboration with Majid Al Futtaim, a company owning shopping malls, hotels, and retail stores across the region. The initial test case involves using Agent Pay for purchasing movie tickets at VOX Cinemas.
Challenges and Considerations
Despite these platforms offering promising use cases, questions persist regarding their effectiveness. For instance, if a user wishes to buy movie tickets, how detailed must their instructions be for the AI agent to meet their expectations? Providing strict guidance on the film and showtime might limit the AI’s usefulness.
On the other hand, many users might prefer retaining control over decisions about evening entertainment purchases. This could result in disputes if customers are dissatisfied with the AI’s choices.
Safeguarding Transactions
Additionally, concerns remain about the security of agentic commerce transactions and the protocols required to prevent fraud.
Broader Acceptance and Adoption
The obstacles facing agentic commerce in achieving wider acceptance include users often wanting final approval before making payments. While many financial services companies have invested significantly in infrastructure for this new paradigm, Visa and Google have also launched protocols to govern AI agents’ actions.
This investment underscores that agentic commerce is not a novelty. Companies should explore how this disruptive technology can shape their operations, even if there is no rush to adopt it.











