UAE Shoppers Open to AI Purchases, But Only 19% of Merchants Ready

What happened

Visa research reported by Arabian Business finds that 82% of surveyed UAE shoppers would allow AI to make purchases on their behalf, provided spending limits are in place. The same findings show that only 19% of merchants have product data that AI systems can read.

Taken together, the figures point to strong consumer appetite for AI-led buying alongside much lower readiness on the retail side. Demand for delegated purchasing appears to be moving faster than the product information needed for an AI shopping assistant to find, compare and buy from a given catalogue.

Why it matters

Agentic commerce changes where the customer experience begins. When a shopper hands a purchase to an AI agent, the agent becomes the first point of evaluation. A product that is poorly described, inconsistently tagged or invisible to machine readers may never reach the consideration set at all. Product information that was written for human browsers now has to be structured for machines as well, which makes data quality a customer experience issue rather than only a technical one.

The spending-limit condition is equally significant. It signals that consumers want control over delegated purchasing, through budgets, guardrails and a clear view of what an agent has bought. For digital transformation leaders, the story is about the operating model. Merchant catalogues, payment rules and service processes all need redesigning for software intermediaries as well as for people.

By the numbers

• 82% of surveyed UAE shoppers would allow AI to make purchases, with spending limits in place.
• 19% of merchants have product data that AI systems can read.

The Renascence take

Most coverage will lead with the 82% figure, because it is the headline most readers will remember. The more useful number is the 19%. Consumer willingness to delegate is rarely the long-term bottleneck; the real constraint is whether a merchant's catalogue can be understood, trusted and acted upon by a machine.

People hand over decisions when they can see the boundaries, which is why the spending limit matters more than the enthusiasm behind it. Legibility and bounded autonomy are the service-design principles underneath this story, and a catalogue that a machine cannot read is effectively a shop with its lights off. Customer-obsessed operators should audit their product data as a customer-facing asset now, and they should design visible confirmation and clear limits before an agent is ever allowed to spend.