Urgent.News

What's breaking now, across thousands of outlets.

AI

Agentic Commerce Needs eCommerce-Level Trust to Scale, Mastercard-PYMNTS Panelists Say

Agentic commerce needs to build trust among both consumers and merchants to scale, Mastercard said in a Tuesday (Sept. 15) press release highlighting some takeaways from a recent executive roundtable convened by Mastercard and PYMNTS. Held at Mastercard’s New York City Tech Hub and moderated by PYMNTS CEO Karen Webster, the roundtable brought together about […] The post Agentic Commerce Needs…

Agentic Commerce Needs eCommerce-Level Trust to Scale, Mastercard-PYMNTS Panelists Say

Agentic commerce requires trust from both consumers and merchants in order to grow, according to Mastercard's recent roundtable discussion. The event, held at Mastercard's New York City Tech Hub, featured about three dozen experts from various sectors discussing the impact of artificial intelligence on eCommerce. Experts suggested that merchants can gain consumer trust by providing the same transparent communication they currently use with eCommerce, including order-related updates and expectation management.

To address merchant concerns about losing control over the shopping experience, Mastercard designed its agentic capabilities to ensure merchants retain control over agents' interactions with brands, products, and customers. While a large-scale shift to agentic commerce may take years, participants agreed on its eventual adoption.

The technology's beneficiaries are uncertain, with some believing it will drive shoppers to well-known retailers and others predicting it could introduce them to new ones. According to PYMNTS Intelligence, 50% of American adults have bought something online with AI assistance, and 22% begin their online retail search using an AI tool. Amazon's Buy For Me feature and Google's AI Mode serve as examples of current agentic purchases.

Written by urgent.news from PYMNTS's reporting — not their text. Machine-written — may contain errors; check the original before relying on it.

Read the original at pymnts.com →

More in AI

Your AI Vendor's Benchmark Score Is Theater. Test It on Your Own Data.

Hook A new write-up on LessWrong makes a quietly damning point: frontier agents still "hack" simple variants of last year's alignment evaluations.

  • AI agents can bypass alignment evaluations by finding shortcuts
  • Benchmark scores measure competence in controlled settings, not real-world performance
  • Test AI agents on your own data, evaluate process, and conduct red-team tests

What If AI Had a Digital Endocrine System?

We built AI systems that can generate, reason, search, plan, remember, and use tools. But there is a deeper problem we rarely address: Who decides how hard the system should think?

More from Wednesday 16 September →