Visa Inc. launched the Visa Agentic Commerce Benchmark on 5 November 2026 in San Francisco, putting 22 AI models from 12 labs through 15 realistic shopping challenges such as buying a television, a laptop or a personal-care appliance. The overall pass rate was 57.2%, with about 43% of runs failing, Visa said in its press release.

Under the benchmark, models must converse with a shopper to uncover hidden buying needs, justify their decision, and refuse to buy if nothing fits. The work sits alongside Visa’s earlier AI payments push on Elets CIO when Visa and OpenAI joined forces for AI-powered commerce, and beside India demand signals such as the McKinsey survey finding 76% of Indian consumers already use AI shopping tools.

What the Visa Agentic Commerce Benchmark measured

Visa framed shopping as hard for AI because buyers often do not know every specification up front. In one television scenario, a customer mentions a relative’s brand tip and a third-party membership but leaves budget, HDR10+ support, format compatibility and gaming needs unsaid. Agents that uncovered more of those hidden factors performed better.

Across all models and scenarios, the five models that uncovered the most buying factors averaged a 70.2% pass rate, versus 44.4% for the five that uncovered the least. The five highest-cost models per run passed 67.1% of attempts, against 50.2% for the five cheapest. Consistency remained weak: all three attempts at a scenario passed only 34.2% of the time.

How top AI shopping agents compared

Best-performing models passed 82.2% of attempts; the weakest passed 28.9%. Visa’s own release did not name individual labs, but Fortune reported that Google Gemini Aragon 4 and Meta Wonder each cleared about 82%, a Claude agent about 76%, and an OpenAI agent about 64%. About 54% of failed runs still ended in a purchase of the wrong product or one that missed stated requirements—a pattern that echoes caution in the TCS enterprise AI study on how few firms let agents act alone.

Why trust still limits agentic commerce

Rubail Birwadker, Global Head of Growth at Visa, said AI agents can power the next era of commerce but that trust is essential. Visa’s Trust Index from September 2026 found 72% of U.S. consumers had used an AI assistant, yet only 23% trusted AI to handle payments on their behalf. That gap is why platforms add governance such as Salesforce Agent Fabric’s kill switch and live evals for AI agents, and why scale debates continue as in GitHub’s view that AI agents could outnumber developers.

Visa Intelligent Commerce—its agentic-commerce portfolio—aims to raise that trust. The benchmark is complementary research, not a product inside that portfolio. Visa plans to extend testing to checkout, payment and order placement and has shared results with model providers. Consumers can try the interactive experience on Visa Payments Frontier. Full methodology is in the Visa Agentic Commerce Benchmark press release.

Key takeaways

• On 5 November 2026 Visa released the first industry Agentic Commerce Benchmark: 22 models from 12 labs, 15 shopping challenges, overall pass rate 57.2%.
• Conversation quality mattered: models uncovering the most hidden buying factors passed 70.2% versus 44.4% for those uncovering the least; all three attempts passed only 34.2% of the time.
• 54% of failed runs still completed a purchase; Visa will extend the benchmark to checkout and payments under its Intelligent Commerce trust agenda.

FAQ

What is the Visa Agentic Commerce Benchmark?

It is Visa’s first industry benchmark for AI shopping agents. A run passes only when the model discovers every task-defined customer requirement and makes a correct purchase or no-purchase decision.

Which AI models scored highest?

Visa did not name models in its release. Fortune reported Google Gemini Aragon 4 and Meta Wonder near 82% pass rates, Claude near 76%, and OpenAI near 64%.

Is the benchmark part of Visa Intelligent Commerce?

No. Visa says Intelligent Commerce is its product and partnership portfolio for trusted agentic commerce; the benchmark is complementary research for the wider industry.

Featured image: Visa Inc. headquarters exterior in San Francisco. Photo: Sarah Stierch / Wikimedia Commons (CC0, August 2025)

Be a part of Elets Collaborative Initiatives. Join Us for Upcoming Events and explore business opportunities. Like us on Facebook , connect with us on LinkedIn and follow us on Twitter.

"Exciting news! Elets technomedia is now on WhatsApp Channels Subscribe today by clicking the link and stay updated with the latest insights!" Click here!