Meridian raises $17 million to remake the agentic spreadsheet

The quest to innovate spreadsheet technology through artificial intelligence continues with the launch of a new player in the field: Meridian. Recently emerging from stealth mode, the company has announced a substantial seed funding round of $17 million, achieving a post-money valuation of $100 million. CEO and co-founder John Ling shared the company's ambitious vision to transform financial modeling, aiming to make spreadsheet processes not only faster but also more reliable and auditable. Ling explained to TechCrunch that their goal is to significantly reduce the time required for financial modeling tasks. "How can you take a process that traditionally might have taken several hours and condense it down into like 10 minutes?" he posed, highlighting the efficiency they seek to offer. This funding round was led by prominent investors including Andreessen Horowitz and the General Partnership, with additional support from QED Investors, FPV Ventures, and Litquidity Ventures. Currently, Meridian is collaborating with teams at Decagon and OffDeal, having secured $5 million in contracts just in December. The focus on AI-driven spreadsheet solutions stems from the high costs associated with manual financial analysis. Unlike previous AI solutions that integrated directly with Excel, Meridian is designed as a standalone workspace, functioning similarly to Cursor. This architecture allows it to combine various data sources and external references seamlessly, reducing potential friction in the workflow. Based in New York, Meridian’s team comprises experts from AI companies like Scale AI and Anthropic, alongside seasoned professionals from financial institutions such as Goldman Sachs. Ling emphasized that one of their main challenges lies in meeting the stringent demands of financial clients, which often conflict with the inherently unpredictable nature of AI models. "If you go to ten different software engineers at Google and request a new feature for an app, you might end up with ten distinct implementations, and that’s acceptable," Ling noted. "However, when asking ten banking analysts at Goldman Sachs for valuation models, you would likely receive ten nearly identical workbooks." To address this, Meridian has invested considerable effort in ensuring their outputs are both auditable and deterministic, while still retaining the adaptable qualities of large language model (LLM)-based tools. Their approach merges agentic AI with traditional tooling, aiming to minimize the inaccuracies that often hinder large-scale enterprise applications. "Our goal is to eliminate uncertainty from the LLM process; you want to understand how the logic flows and where all assumptions in the model originate," Ling concluded.