Ramp, the leading financial operations platform, announced Agents for AP, specialized AI agents to automate the most manual and time consuming parts of accounts payable: invoice coding, invoice approval and payment processing. Agents for AP work directly in Bill Pay, Ramp's AP automation platform with industry leading OCR. Bill Pay can process, approve, and pay invoices in as few as three clicks.

"Ramp's Bill Pay was already automating the majority of the AP process, and Agents for AP are now handling more complex workflows — nearing 100% automation in certain cases," said Karim Atiyeh, co-founder and CTO at Ramp. "This is only possible with agents that can source, understand and apply years of contextual knowledge to make decisions on crucial details like GL codes and approval recommendations."

Agentic Workflow Automation for Accounts Payable 
Agents for AP use the same context finance teams and approval stakeholders rely on to make decisions autonomously, connecting each invoice with vendor records, contracts, purchase orders, and approval history. Agents for AP can:

• Code invoices – Agents learn and apply logic from historical data and invoice details, like product IDs and shipping addresses, to input general ledger codes for each line item. Agents for AP now handle most invoice coding tasks automatically, applying learned logic to get 85% of accounting fields right the first time and continuously improving with every cycle.
  
  
• Streamline approvals with recommendations – Agents provide approval recommendations and a summary of vendor history, contracts, prior bills, and coding consistency, giving approvers all the information to make a decision without searching.
  
  
• Apply card payments – Agents find card payment opportunities directly in the vendor's payment portal, eliminating manual entry and capturing cashback.

Andrew Clarke, VP of Finance at STUDS: "For the finance team, Ramp has been a true game-changer. We can instantly analyze spend by vendor across all payment methods, and with the introduction of Agents and AI coding, we've eliminated tedious manual work while streamlining internal processes."