What it is
Agent commerce describes the set of business activities that emerge once agents can spend money autonomously within budgets and policies their owners define. Concrete examples: an agent paying $0.02 to a paid weather API to answer a logistics question, an agent buying 30 minutes of access to a research database to produce a report, an agent hiring a specialized "translation agent" through a marketplace and paying it on completion, two agents negotiating a freight rate and settling instantly. The category requires three substrate pieces working together: a payment protocol that supports machine-speed micropayments (L402 over Lightning, x402 over USDC, or similar), a discovery mechanism that lets agents find paid services with structured pricing metadata, and an authorization layer that binds an agent's spending to its principal's budgets and policies. As of 2026, the substrate is maturing rapidly even though the volume is still small.
Why it matters
Most of the value of "AI agents replacing knowledge work" is gated behind whether agents can transact. An agent that can read paid content, license premium data, and pay other agents for specialized work is meaningfully more useful than one that can only call free APIs. Once the substrate is in place, two structural effects follow: (1) the long tail of specialized data and capability becomes economically reachable for the first time — small publishers and tool authors can charge fractions of a cent per call where credit cards required dollars; and (2) entire businesses become possible whose customers are agents rather than humans. For Salesforce-centric organizations in 2026, the practical near-term question is whether Agentforce agents will participate in agent commerce through the Agentforce Exchange or through open protocols, and what that means for procurement and finance controls.
Key components
- Machine-payable APIs — services that expose their price natively in HTTP responses for autonomous payment
- Payment protocols — L402 (Lightning), x402 (USDC), and emerging successors
- Discovery + pricing metadata — structured ways for agents to find paid services and understand their cost
- Spending authorization — budgets, policies, and approval flows binding agent spend to principal accounts
- Settlement and reconciliation — the bookkeeping layer that ties machine-speed transactions back to the org ledger
Related terms
AI Agent
An autonomous AI system that can perceive its environment, make decisions, and take actions to achieve specific goals - without constant human direction.
Agent Operations
The discipline of running AI agents in production — capturing what they do, attributing what it costs, evaluating what they produce, and intervening when something goes wrong. The operational layer above agent observability and orchestration.
Agent Infrastructure
The runtime, network, and tooling substrate that AI agents need to execute reliably — sandboxed compute, tool access, memory, gateways to LLM providers, and the orchestration plumbing that connects them. Closer to the metal than agent operations.
L402 Protocol
A machine-to-machine payment protocol that combines HTTP status code 402 ("Payment Required") with the Bitcoin Lightning Network, designed for AI agents and APIs to transact in fractions-of-a-cent without involving credit cards or human approval. Originally called LSAT by Lightning Labs.
Machine-Payable APIs
APIs that expose their price natively in their HTTP response and accept payment from a calling client without human involvement — the substrate that makes agent commerce possible. Typically implemented over HTTP 402 plus a payment protocol like L402 or x402.