Opening This chapter synthesizes the channel's material on agentic finance, focusing on payments and accounts built for AI agents. As software agents move from answering questions to taking actions on behalf of users, the financial infrastructure they rely on must evolve alongside them. The core question addressed across the material is how agents can hold, move, and spend money in a way that is safe, accountable, and compatible with existing financial rails. Sections The discussion first frames what agentic finance means: a shift in who initiates financial transactions. Traditionally, a human clicks the checkout button or authorizes a wire; in an agentic model, the agent itself becomes the transacting party, operating under delegated authority from a person or organization. This reframing drives new requirements for identity, authorization, and control at the payment layer. The material then turns to accounts for AI agents. Giving an agent its own account — or a scoped sub-account within an existing account — allows spending to be bounded, monitored, and attributed to the agent rather than to the human owner. Separation of this kind supports auditing, budgeting, and revocation, which are essential when autonomous software can initiate real economic activity. Payments are covered next. Agent-initiated payments need mechanisms that let merchants and payment networks recognize the agent as acting on behalf of a verified principal. Delegated credentials, per-transaction limits, and machine-readable authorization are presented as the building blocks that let agents complete purchases, subscriptions, and disbursements while keeping humans in the loop where it matters. Key Points Agents transacting directly changes the trust model: authorization and attribution must move into the payment stack itself. Dedicated or scoped accounts give organizations control over what an agent may spend, making agent activity auditable and revocable. Payments for agents rely on delegated identity and limits so that merchants, networks, and regulators can trust machine-initiated transactions. The goal is not to replace existing rails but to extend them so agents can operate within familiar financial infrastructure. Closing Taken together, the material presents agentic finance as an infrastructure problem: payments and accounts must be redesigned around delegated, machine-initiated activity without sacrificing the control, visibility, and trust that human-centered finance assumes. Organizations that adopt agent-native accounts and payment controls can let agents act with real purchasing power while keeping every action attributable and bounded.