AI Agents Will Decide Which Brands Win—Here’s What They’re Looking For

Forbes
“Brand is a data problem now, and most brand teams aren’t staffed or structured to treat it that way.”
AI agents are becoming the interface between brands and customers. In this Forbes Agency Council article, Fantasy CEO Firdosh Tangri argues that agents choose brands on transparent, consistent, machine-readable information, and that decades of accumulated brand goodwill no longer decide who gets recommended.
The stakes are large. Morgan Stanley projects that AI agents could drive $190 billion to $385 billion in U.S. e-commerce spending by 2030.
Key points
How do AI agents decide which brands to recommend?
- Agents compare the data. They weigh brands on data quality, pricing consistency and transparency. If product data is messy, agents recommend a competitor.
- Design for machines. Structure product information, policies and service capabilities so agents can parse and compare them accurately.
- Keep data consistent. A price discrepancy, a conflicting description or outdated inventory anywhere an agent meets the brand erodes trust.
- Set policy for your own agent. Decide what it can promise, what it can discount and when it hands off to a human.
- Earn the trust layer. Getting design, data and policy right together is what earns the recommendation.
Frequently asked questions
Tangri describes four modes of purchasing. Mode one is the fully human purchase decision. Mode two is people assisted by agents for product research, comparison shopping and deal finding. Mode three is the agent mediating the purchase with human approval at the end. Mode four is agent-to-agent transactions with no human in the loop.
Not fully, yet. Research from the Interactive Advertising Bureau found that 89% of consumers double-check what an AI agent recommends before buying. Tangri sees that gap between convenience and confidence as where the next decade of brand competition will play out.