Ramp adds controls for enterprise AI token spending

Ramp Business Corporation has expanded its AI Token Spend Management product, adding a consolidated view and controls for usage-based AI costs across multiple model and developer-tool providers. The company did not disclose pricing for the expanded product beyond saying it is free to start for Ramp customers and non-customers.

The move puts Ramp deeper into a cost category that has grown quickly as engineering, product and operations teams connect applications to large language models. For finance teams, the problem is less whether AI budgets are rising than whether anyone can tie invoices back to teams, projects, API keys and actual consumption before the bill arrives.

Ramp said the product can ingest spending from providers including OpenAI Group PBC, Anthropic PBC and Google LLC’s Gemini. It also supports connections to Cursor and OpenRouter, according to the company. Once connected, customers can review spend by provider, team, employee, project or API key.

The expanded system includes weekly spending briefings, reconciliation between invoices and usage, and alerts intended to flag overruns before they become finalized costs. Finance teams can also set limits at the team, project or API-key level and send anomaly alerts to the relevant stakeholders, Ramp said.

Why AI spend is harder to manage

Ramp’s argument is that token usage does not fit cleanly into the traditional corporate spend buckets of headcount and vendor contracts. AI costs are metered, can be generated by many teams at once and often sit behind API keys rather than named cardholders. Invoices may show the vendor charge without enough detail for finance to understand which application or team created the spend.

That creates an opening for spend-management vendors. Ramp is positioning the product as a way to bring AI usage into the same finance workflow as other operating expenses, rather than leaving model bills to engineering dashboards or manual spreadsheet work.

Ramp said thousands of companies have connected AI providers to its platform, though it did not provide a precise customer count for the token product. The company also said the system often identifies cheaper model options for the same workload. According to Ramp, one-third of customers have access to a lower-cost model that can perform the same task as effectively, but it did not disclose the data set, workloads or evaluation criteria behind that claim.

The company cited AngelList Advisors LLC as one customer example. Ramp said a weekly briefing pointed AngelList to prompt caching after the company had been losing $10,000 a month, and that engineering applied the change the same day. Ramp did not state the resulting savings.

Ramp’s own AI usage helped shape the product

Ramp said the offering originated from internal tooling built for its finance organization. The company claims 99.5% of its employees use AI tools daily and that token usage has become a meaningful operating expense inside Ramp. It did not disclose total AI spend, revenue contribution from the product or expected adoption targets.

The launch comes shortly after a large financing round for the New York-based fintech company. Founded in 2019, Ramp raised $750 million in a Series F round in June at a $44 billion valuation. ICONIQ Capital, GIC Pvt. Ltd. and Ontario Teachers’ Pension Plan led the financing, which brought Ramp’s total funding to more than $3 billion.

For Ramp, AI token management is a logical extension of its core pitch to finance teams: centralize spend data, add policy controls and reduce waste. The open question is whether AI usage becomes a durable software category for finance departments or remains a feature absorbed into broader procurement, cloud-cost and developer-platform tools.

This story draws on original reporting from SiliconANGLE.