Observed Signal · Nov 18, 2026 · Technical Release · Source: Linas Newsletter · Impact: 4/5 · Sentiment: Positive

Visa Benchmark: AI Shopping Agents Fail 23% of Purchases

Executive Signal Summary

Visa tested 22 AI shopping agents built on models from Google, Meta, OpenAI, Anthropic, and eight other labs. In 23.1% of runs, the agent bought a product that did not meet the shopper's needs, leading to potential disputes. Visa's Agentic Commerce Benchmark highlights that even top models fail and that a more expensive model doesn't solve consistency issues. The company aims to establish itself in the agentic payments landscape, where liability for agent-initiated purchases remains unresolved. The test suggests a significant financial impact from disputes, positioning Visa as a potential arbiter in this emerging space.

Polaris7 AgentPolaris7 Strategic Assessment
High Confidence

Visa's benchmark provides crucial data on AI agent reliability, highlighting liability and dispute risks that will shape the future of agentic commerce and payments.

MARKET SIGNALS

Track Visa Signals & Strategic Shifts

Polaris7 connects regulatory filings, primary sources, leadership changes, and M&A activity into structured intelligence. Set up your Polaris7 workspace to monitor relevant companies in your executive briefing.

  • Verified primary sources
  • Automated briefings
  • Instant watchlist setup

Key Takeaways & Evidence Grounding

  • Visa tested 22 AI shopping agents built on models from Google, Meta, OpenAI, Anthropic, and eight other labs.
  • In 23.1% of runs, the AI agent purchased a product that did not meet the shopper's needs.
  • Visa's Agentic Commerce Benchmark reported a 57.2% pass rate for the agents.
  • The article highlights that no party has yet agreed on liability when a shopper disputes an agent-initiated transaction.
  • Visa aims to position itself as a central figure in the agentic payments disputes landscape.

Connected Companies & Entities

Ontology Mapping & Concepts

Agentic Advertising & Autonomous Media (ADCP)#Agentic Commerce Benchmark#AI Shopping Agents
Primary Source Grounding & Direct Attribution
Direct Origin Attribution
Primary Reporting: Linas Newsletter•Published: Nov 18, 2026
Original Coverage Title: “Visa Tested AI Shopping Agents. 23% Bought the Wrong Thing”

Related Market Signals & Shifts

Recent verified developments and strategic activity across this market segment.

Agentic CommerceSep 11, 2026

Visa Study: Trust in Established Payment Brands Key for Agentic Commerce

A new Visa study reveals that while 72% of US consumers have used an AI assistant, only 23% trust generative AI to independently handle payments. However, trust increases significantly when established payment brands are involved, with 61% willing to entrust Visa with agentic transactions, rising to 68% among 18-34 year-olds and 71% among frequent AI users. The study, conducted by Harris Poll in late May 2026 with 2,065 US consumers, highlights a distinction between trust in AI for product search and the need for reliable payment infrastructure. Visa is developing 'Visa Intelligent Commerce' to build technologies and standards for AI-agent-initiated transactions, focusing on identity verification, authentication, and consumer controls. The findings suggest that agentic commerce is still in its early stages, with a significant acceptance gap between using AI assistants and fully autonomous purchases.

Read assessment
Agentic CommerceSep 3, 2026

Investors Are Betting on AI Agents That Shop for You

Venture capitalists are increasingly funding startups building AI agents that can shop on behalf of consumers and handle payments. The article maps the emerging market for "agentic commerce," covering personal assistants like Instinct and Town, shopping search tools like Daydream, payment infrastructure startups like Catena Labs and Basis Theory, and marketing tech such as Profound. Major platforms are also moving: OpenAI and Stripe launched the Agentic Commerce Protocol, Google operates the Agent Payments Protocol, and Visa and Mastercard introduced Trusted Agent Protocol and Agent Pay. However, the space faces trust and fraud challenges, illustrated by Phia's cookie-stuffing scandal and an Instinct agent error that cost a user around $300. Investor interest remains strong, with millions raised across the sector, even as enterprise adoption lags due to compliance and security concerns.

Read assessment
Agentic CommerceSep 2, 2026

Mastercard Research Highlights AI Agents as New Brand Challenge

Mastercard's new Signals report, 'Encoding Trust: The Race for Intent, Consent and Control in the Agentic World', examines how AI agents will reshape commerce and what brands must do to win their trust. It forecasts agent-assisted consumer spending will reach $3-5 trillion by 2030, with B2B spending potentially three times higher. Survey data shows 85% of consumers are open to using AI agents to find products, 74% are willing to delegate specific commerce tasks, but only 10% would let an agent complete a purchase fully autonomously. Mastercard argues that scale depends on consent-based, visible and reversible transactions, proposing 'agentic tokens' and a trust layer to limit agents' spending powers. The report also warns that traditional human persuasion tactics, such as scarcity messaging and FOMO, are ineffective with AI agents, so brands must prioritize accurate product data, competitive pricing and authentic ratings.

Read assessment

WATCHLIST & BRIEFINGS

Receive future signals for Visa & competitors in your briefing

Stay ahead of material M&A deals, leadership changes, account wins, and product rollouts. Polaris7 connects primary filings, company newsrooms, and trade reporting directly into your executive briefing.

  • Regular primary-source checks
  • Structured Watchlist briefings
  • Preserves current context