Agentic AI Platform Change Log

Agent capability

Ramp for Nonprofits includes Grant Policy Agent, which checks proposed expenses and bills against a grant's spending policy and flags likely mismatches before approval or allocation, such as a hotel bill against a grant that excludes travel. The launch also adds Grants Hub, grant budgets with spent, committed and remaining balances, program cards tied to a grant, and sales tax reclamation.

Checking spend against grant terms is manual work that small finance teams often skip, and an agent that screens each request before a person approves it lowers the risk of misallocating restricted funds.

Agent capability

Ramp launched Accounts Receivable with agents that draft invoices from contracts, prepare collection messages using customer history and policy, and match incoming payments to open invoices. Finance reviews invoice drafts and sends every collection email, while unmatched payments are flagged for review. The help center confirms live core invoicing for US businesses using QuickBooks Online; NetSuite access is still subject to rollout and a waitlist.

Finance teams can bring collections and cash application into their existing Ramp workflow while retaining control of customer communications. Buyers should verify accounting provider eligibility and distinguish free core invoicing from the paid billing schedules and revenue recognition plan.

Observability / auditability

Ramp's audit log events now include agent attribution, with an agent_name field whenever an AI agent acts on a user's behalf.

Finance teams need to know whether a person or an agent approved a payment. Agent attribution in the audit log is the minimum control for that.

MCP / tool calling / API

Ramp launched Router.com, a unified API designed to route corporate AI requests to the lowest-cost models available. The service dynamically selects the most cost-effective model for each request to reduce overall AI inference costs.

Engineering and finance teams can integrate this API to automatically optimize their AI spending without manually switching between different model providers. This allows companies to scale their AI agent and application deployments while maintaining predictable inference budgets.

Agent capability

Ramp launched Ramp Router in closed beta, opening its internal large language model gateway to developers. The OpenAI-compatible API dynamically routes AI requests across a catalog of models and service tiers. It uses Thompson sampling to optimize for latency, failure rates, and cost, allowing teams to switch models without rewriting application code.

Engineering and finance teams can manage AI infrastructure centrally to reduce token spend and prevent vendor lock-in. By offloading model selection to a dynamic router, organizations ensure they use the most efficient model per request while maintaining full visibility into per-request costs.

MCP / tool calling / API

Ramp introduced Ramp for Agents, a feature that enables users to incorporate a business, apply for corporate cards, and establish a finance stack through a single text prompt. This release includes a Model Context Protocol (MCP) integration and a CLI toolset with over 50 finance playbooks, allowing AI agents to perform tasks related to banking, bill pay, travel, expenses, and accounting.

Founders can automate administrative setup tasks without leaving their development environments. The provided MCP toolset allows engineering teams to programmatically embed financial operations and autonomous spending controls directly into their custom internal workflows.