Synthetic Aims for Autonomous Accounting Agent, Integrating AI with Core Financial Systems

Ian Crosby’s vision for Synthetic is ambitious: to build a truly autonomous AI agent capable of managing the entire bookkeeping lifecycle. This isn't merely an assistive tool that prompts human intervention; it aims to function as a full operational replacement for manual financial close processes. The system connects directly to disparate sources—banking feeds, payroll systems, billing platforms, and email inboxes—to process transactions automatically. The core ingenuity lies in the architectural shift from 'assistive AI' to 'autonomous execution.' Crosby frames this by comparing it to advanced self-driving technology, asserting that responsibility for outcomes resides entirely within the system itself. This requires the AI agent not only to perform calculations and classifications but also to determine what information is missing and formulate precise, clarifying questions—a sophisticated process of identifying gaps in data rather than simply flagging discrepancies. This vision pushes against the known limits of current enterprise resource planning (ERP) and financial close automation platforms. Competitors like Optimus focus on automating the reconciliation and journal entry phases, improving speed and compliance through structured workflows across existing systems. Synthetic seeks to bypass that structure altogether. While modern finance tools centralize visibility and automate matching across ledgers and bank feeds, Synthetic attempts to consolidate both the data ingestion layer (from emails/inboxes) and the process execution layer into one self-correcting loop. Given Crosby's background—co-founding Bench Accounting and Teal, with a deep understanding of accounting pain points—the company has established significant market credibility. The $10 million seed round, backed by investors like Khosla Ventures and Shopify’s Tobi Lütke, reflects confidence in his ability to navigate the inherent high risk of this venture. Achieving 100% accuracy in bookkeeping is a massive technical hurdle, but if successful, Synthetic would consolidate multiple stages—from transaction capture (the messy inputs) to reconciliation and journal entry (the clean outputs)—into a single, autonomous loop. It represents an attempt to solve not just how accounts are closed quickly, but how the data for those accounts is generated and understood in real-time. For Canada, this focus on deep fintech infrastructure signals a sophisticated maturity of local tech talent. While Synthetic operates from San Francisco due to cultural considerations, the problem domain—Canadian SMEs struggling with bookkeeping complexity—remains profoundly relevant. An operationalized system like Synthetic would radically simplify compliance and cash flow management for Canadian businesses, making complex financial operations accessible via simple button pushes.