Ramp Debuts Stack, an AI Accounting Operating System

Ramp has made its most aggressive product bet yet. On June 3, 2026, the $32 billion financial operations company launched Stack, an AI accounting operating system that puts autonomous agents to work on the grunt labour of professional accountancy: bank reconciliations, journal entries, depreciation schedules, prepaid amortisation, deferred revenue roll-forwards, variance analysis, month-end closes. Not one of those tasks in isolation. All of them, end to end, across an entire client book. The target market is the roughly $150 billion US accounting firm sector, and the pitch is blunt. Stack does not assist accountants. It does the accounting, under their supervision, and logs every step for audit.

The timing is not accidental. More than 300,000 CPAs have left the profession in recent years. Accounting degree completions are at a 20-year low. Firms across the US are sitting on full client pipelines and half-empty offices, turning away revenue because they cannot hire fast enough to service it. The businesses they serve are not waiting around either. They want faster closes, real-time financial visibility, and advisory that goes beyond compliance. The firms that cannot deliver that are losing mandates to the ones that can.

Every technology company in finance has noticed the opportunity. Microsoft has pushed Copilot into Dynamics. Intuit is threading AI through QuickBooks. Dozens of vertical SaaS startups are chasing audit prep, tax automation, and bookkeeping. Wolters Kluwer recently expanded its enterprise collaboration with OpenAI. But most of these products still treat AI as a feature layered onto existing software, a copilot that suggests rather than executes. Ramp is arguing that the industry needs something fundamentally different: a standalone AI accounting operating system built from scratch on real close data, not a general-purpose model dressed in accounting clothes.

Firms connect a client’s systems, upload their close checklist, and define how they want the work done. The platform codifies those processes into what Ramp calls “Skills,” editable standard operating procedures that run identically across every client, every month. As the firm refines its approach or a client’s business changes, the Skills update. The institutional knowledge stays inside the firm as proprietary IP.

“Accounting firms are under more pressure today than at any point in history. The firms we work with aren’t asking for another AI tool to prompt. They need something that actually does the work, with every decision reviewable and auditable.”

Geoff Charles, Chief Product Officer, Ramp

Ramp published benchmark results from more than 200 accounting tasks it says were built and graded by practising accountants, covering transaction coding, reconciliation, journal posting, and full close execution, the AI accounting operating system beat ChatGPT, Claude, and Gemini by four to eleven percentage points. Domain-specific training is winning over raw model power, at least for now.

The early numbers from live firms are harder to dismiss than benchmarks. Ramp says Stack can cut month-end close times by up to 50% and run recurring schedules nine times faster than manual methods. Firms that helped build the product are already seeing it.

“Stack is eliminating or reducing the amount of time we spend on a month-end close by 50% on some clients. My role is shifting toward people investment. The tools are here to systemise the work.”

Tyler Otto, President and Owner, Specialized Accounting

Other early users described a more fundamental shift in how their firms operate.

“We’re running multiple agents simultaneously, spending time reviewing work, not doing and then reviewing. It fundamentally changes the way we work.”

John Ikosipentarhos, President, Zeroed-In Consulting

Everything runs through a centralised dashboard called the Advisor Console, where firm staff monitor agent work in real time, inspect the reasoning behind each output, and approve anything that requires professional judgment before it posts. The audit trail is granular: every source document, every calculation, every decision point is traceable back to its origin. The platform plugs into the systems firms already use: QuickBooks, bank feeds, Google Drive, spreadsheets. No platform migration required. Ramp says a new client can be onboarded and running within an hour, a claim aimed directly at smaller and mid-market practices that have historically lacked the IT capacity to implement enterprise-grade software.

On data privacy, Ramp is making promises it knows the profession will hold it to. The AI accounting operating system runs on SOC 2 controls with encryption in transit and at rest. Workflows and firm IP stay walled off from other firms, from clients, and from the AI labs whose foundation models power the system. Contracts with those providers include zero data retention. For an industry where a single data breach can end a client relationship, those are baseline requirements, not selling points.

Ramp has distribution that most competitors in this space simply do not. The company partners with more than 4,500 accounting firms already, many of which served as design partners for Stack during development. Ninety-two of the top 100 CPA firms have clients on its financial operations platform. It serves over 50,000 businesses, processes more than $100 billion in annual purchase volume, and raised $1.15 billion in 2025 alone, pushing its valuation to $32 billion. Reports in May 2026 pointed to a further $750 million raise at north of $40 billion.

The flywheel writes itself. The more firms that adopt this AI accounting operating system, the more of their corporate clients land inside Ramp’s broader ecosystem of cards, expense management, and accounts payable. The accounting firm becomes the acquisition channel.

Stack is free through August 2026 and available now at ramp.com/stack.

The question that matters is not whether the AI accounting operating system works in a demo or a controlled benchmark. It is whether it holds up when a firm onboards a client with ten years of neglected books, three overlapping charts of accounts, and a carrier bag of unreconciled receipts. That is where automation ambitions in this profession have historically gone to die. Ramp has the money, the reach, and the early converts. Now it has to prove it can close the books.