Ramp releases AI agents for expense approvals, analytics

Financial operations platform Ramp announced the release of its first set of AI agents, which the company said are "built and trained to think like your sharpest controller and work like a thousand of them—around the clock."

These agents, meant primarily for controllers, apply context-aware, human-like reasoning to manage entire workflows independently and proactively through their ability to reason and act on behalf of team members. New users will upload a PDF outlining their policies, and the agent builds a "reasoning graph" in minutes that will then allow it to make decisions and learn in real time. If a user does not have a policy, the agent can help craft one. 

Once the agents know the policy, they can autonomously approve low-risk expenses or provide a recommendation with a rationale to the approver, though they will also alert the user in the event of suspicious receipts and invoices. Further, the agents can answer employee questions about their own company's spend policies, responding to messages directly over Slack, email or text. They will also provide data analytics, such as uncovering trends that might signal fraud or waste. 

The AI agents actively learn not just from the company policies but from user activity. By observing how staff actually operate in the system, as well as receiving direct feedback from them, the AI can suggest edits and improvements to those policies. Users can also monitor the agents themselves: the solution lets people see how well an individual agent is performing and how people respond to it (do they ignore it, disagree with it, listen to it, etc.), which, in turn, serves to train the agents further. 

"Ramp agents have complete knowledge of your accounting rules and expense policies that employees don't carry in their heads, plus instant access to transaction details that finance teams would need time to gather. This lets them act faster and more accurately on every transaction," said Karim Atiyeh, co-founder and CTO at Ramp. "This isn't just automation. It's intelligent reasoning that handles complex financial decisions to reduce errors, strengthen policy enforcement and stop fraud."

Ramp says this is just the first step in its larger goal of building a fully autonomous finance operations stack from purchase order requests to reconciliation during month-end close to budgeting agents. Over the coming months, Ramp plans to release agents that handle procurement, intake and vendor onboarding as well as make purchases on users' behalf; reconciliation agents that auto-match transactions to books, flag mismatches, and reduce close time; and budgeting and reporting agents that track against plans in real time and surface anomalies before they hit review.

The news comes a few months after Ramp, in response to reports that AI is now capable of generating convincing-looking financial documentation, designed and implemented a tool to detect fakes. Rather than try to assess the image itself, the software looks at the file's metadata for markers specific to generative AI systems. Once those markers are present, the software flags the receipt as a probable fake.