Datarails, a company that makes financial planning and analysis software, is making a bold bet that the traditional FP&A tools it helped pioneer are now obsolete thanks to AI, and that it needs to disrupt itself before someone else does.
In response, Datarails is launching FinanceOS, an AI-native platform it describes as a “financial operating system”—a platform that allows finance teams use whichever AI tools they want, like Anthropic’s Claude, OpenAI’s ChatGPT and Microsoft Copilot, to perform financial analysis, while maintaining necessary data controls and audit trails.
“AI can build models and run analysis and create reports much faster and much better than any human,” Didi Gurfinkel, the company’s cofounder and CEO, told Fortune in an interview. “So all these tools that focused on creating tools for people, for humans—they’re not relevant anymore. The opposite. They limit the AI.”
It’s a provocative claim from a decade-old company that made its name solving what Gurfinkel calls “Excel hell”—the challenge of managing the sprawl of spreadsheets that finance departments rely on for budgeting, forecasting, and reporting. Datarails built a platform that consolidated data from accounting systems, HR platforms, CRMs, and other operational software into a single source of truth, then connected that data to the Excel models that finance teams already used. Datarails, which is based in Tel Aviv, Israel, has raised $175 million in venture capital funding to date, including a $70 million Series C funding round in January.
But the arrival of generative AI, Gurfinkel said, has changed what’s possible—and what’s needed. AI models can generate sophisticated financial analyses in seconds, but chief financial officers can’t simply throw their data into ChatGPT or Claude and trust the output.
“The one challenge or problem that currently CFOs have with AI is trust,” Gurfinkel said. He breaks this into two dimensions: trusting the data the AI is working with and trusting that the AI’s output is repeatable. The latter is especially challenging since the leading AI models are inherently probabilistic and won’t give the exact same answer to the same prompt every time.
Datarails hopes to address both of these issues with its new FinanceOS product. The system connects data from more than 400 different sources—the “systems of record” that finance teams rely on, such as NetSuite, SAP or Salesforce—and then performs real-time financial consolidation of this data, including complex eliminations, allocations, and foreign exchange adjustments. The platform then lets AI models analyze this data using Model Context Protocol (MCP), the emerging open standard for connecting AI systems to external data sources.
Then, once a financial model is built with AI, FinanceOS allows a customer to lock that model in place so that the finance model remains consistent, while the underlying data refreshes each period.
Datarails’ timing may be right. According to a Gartner survey cited by the company, AI adoption in corporate finance functions has essentially flatlined, rising just one percentage point, from 58% in 2024 to 59% in 2025, while 91% of finance teams report low impact from their AI tools. Data quality and availability were cited as the most common obstacles.
At a time when investors are hyper-focused on how AI challenges the traditional license payment per user business model from software-as-a-service vendors, Datarails is leaning into disruption. It’s shifting to a usage-based pricing model, which Gurfinkel said makes sense as AI agents, not humans, are increasingly using software.
“Total spend on software will be higher—it will increase,” he said. “But probably the number of people will be less. AI can do more. So if you take this equation, you get to one very obvious conclusion: the CFO will pay by the value.” Gurfinkel said that usage-based pricing is a proxy for the value a company derives from using a product.
Datarails is positioning itself not just as a product company, but also as a partner helping CFOs navigate the AI transition. Alongside FinanceOS, the company plans to offer professional services, training, and custom agent development—an acknowledgment that, as Gurfinkel puts it, “the office of the CFO is the last to adapt new technology.”
This hands-on approach echoes the strategy being pursued by other companies selling AI agent-based products to enterprises, including Salesforce, Anthropic, and OpenAI, which have recruited teams of “forward-deployed engineers” that help customers design agentic workflows and configure AI systems, as opposed to the older model for SaaS companies which was largely about self-service by customers.
Gurfinkel was blunt about the competitive landscape, arguing that many of the industry’s oldest FP&A software vendors are in trouble. “They’re already gone. They are slow. They don’t have enough cash or energy to rewrite the technology,” he said. Newer entrants such as Abacum and Runway, which invested heavily in sophisticated web interfaces and algorithmic workflows, face a different challenge: They need to reinvent themselves after underinvesting in the data consolidation layer that Gurfinkel believes is the new strategic high ground.
The good news for those companies, he said, is that most have recently raised significant capital, giving them time to adapt. “But it will be interesting to see how they react to this change,” he added.
He draws a parallel between what he predicts will happen to financial professionals and what is already happening in software engineering, where AI coding assistants have transformed how developers work. “You don’t see any programmer that actually types on their keyboard,” he said. “Almost 100% of their code is written by AI. And I’m confident that it will be exactly the same for finance people.”
Datarails said FinanceOS is available immediately and can be fully operational within a few business days, the company says. Datarails’ existing FP&A, cash management, month-end close, and spend control products remain available as managed solutions built on the same underlying platform.
