Will brands succeed by building their own AI shopping agents?

Accenture argues that once AI agents mediate purchasing, "the moment of choice moves upstream to the decision layer—the standards, data and verification logic that determine what gets chosen," and that brands face a binary: "become the agent of choice, or be chosen by the agent." The firm reports that 71% of consumers "expect generative AI to influence at least half of their spending decisions over the next 12 months," and that agents "will break brands down into comparable parts, test claims against reality and steer spending toward the best outcomes," so that "weak differentiation, inflated pricing and poor fit get exposed instantly." Specific appetite is concrete: 61% want an agent that shops across multiple grocery stores, and 71% want one that plans "a complete trip across airlines, hotels and activities." Accenture adds that weekly gen-AI users are "more than twice as likely to want retailers or brands to become the agent of choice than AI-native platforms."

The mechanism Accenture gives is structural, not just behavioral: when an agent sits between consumer and brand, it reorders "what gets bought, when, and how brand value is created," so treating agents as "just another channel" misreads what's happening. Accenture qualifies this against pure commoditization — "nearly 63% want agents to shop for their 'idealized self'" — so a brand with real differentiation is amplified by an agent's comparison, not just exposed by it.

Accenture's prescription — that brands should build their own agent to become "the agent of choice" — sits in tension with Can an AI agent serve both merchant and user interests fairly?, which argues a brand- or merchant-funded agent cannot represent the user's interest without conflict once it earns referral fees from the merchants it recommends. Accenture's own figure that weekly users prefer retailer-built agents over AI-native platforms describes a stated consumer preference; it does not address Azhar's structural objection to who an agent answers to once built. The claim that agents will "test claims against reality" also depends on exactly the ecosystem conditions in Why do capable AI agents still fail in real deployments? — verification logic requires standardization and trustworthiness that current agents are not shown here to have.

This excerpt does not establish that agents are currently performing this claims-testing and comparison function at the sophistication Accenture describes; the figures are consumer expectations and preferences reported to a survey, not measurements of agent behavior or market outcomes. Accenture also sells agentic-AI transformation strategy to the brand audience this report targets, giving it a direct interest in framing the shift as urgent and binary ("the window to decide where to play is open" but "habits forming now will be hard to break"). The reasonable inference is that consumers expect and would welcome agent-mediated comparison shopping, not that this comparison infrastructure already functions as described.