The Rise of Agentic Commerce: When Machines Become Buyers

Trade and commerce have always been vital to society, evolving from barter systems to today's complex digital ecosystem. In recent years, the online growth of e-commerce has skyrocketed, with a market value of $7.4 trillion, projected to reach $8 trillion by 2027. As technology advances, AI-powered agents, or "agentic AI," are taking on more tasks, leading to the emergence of agentic commerce.

This new form of e-commerce allows autonomous AI agents to research, compare, negotiate, and complete purchases on behalf of users or businesses, often without direct human involvement.

For agentic commerce to function, it needs structured product or service data with standardized protocols. This enables the agent to discover items, verify user authorization, and execute payments safely. Users can delegate purchasing tasks to agentic AI, including buying multiple items or services, such as flight tickets, local transportation, and hotel bookings.

Agentic commerce shifts the buying journey from human-driven to AI-driven. Instead of spending time on manual searches and comparisons, users simply provide intent (e.g., "find a business class ticket under 2,000 dollars") and the AI agent handles the entire process. Agentic commerce can be applied to various use cases, like recurring purchases, B2B procurement, price monitoring, travel, and digital services.

The four-part flow of agentic commerce is:

1. Intent capture: Users set constraints for the agent to act on (type of products/services, budget, size, timing).

2. Discovery and selection: Agents scan multiple service providers using machine-readable product data across various sites.

3. Authorization: Agents use cryptographic proof to ensure the user approves specific purchases.

4. Payment completion: Agents use secure tokens to finalize payments, and merchants fulfill orders during checkout and settlement.

Key characteristics of agentic commerce include autonomous decision-making, end-to-end execution, reasoning and planning capabilities, and interoperability. Benefits of this emerging commerce model include personalization, faster purchasing, lower transaction costs, improved decision-making, and seamless shopping experiences.

However, agentic commerce also faces challenges, such as the principal-agent problem. When AI agents act on behalf of consumers or firms, trust issues arise – can users trust that their agents will act in their best interests? This requires mechanisms for alignment, transparency, accountability, consent, and oversight to ensure AI agents genuinely represent the principal's interests. Other challenges include privacy, data security, and alignment with the user's actual preferences.