In a corner of the software world where trust matters more than speed, a small team in Bulgaria is attempting something audacious—teaching artificial intelligence to do the kind of accounting work that could hold up under the scrutiny of a Big Four audit.
Quillon, which just secured $1.5 million in pre-seed funding, is building AI tools for technical accountants, the specialists who navigate the thorniest questions in financial reporting: how to account for a complex debt restructuring, whether revenue from a multi-year software contract should be recognized now or later, what disclosures an M&A transaction demands. The April 24 funding round was led by 42CAP, a Munich-based firm focused on early-stage B2B software, with backing from undisclosed angels affiliated with NVIDIA and Roblox.
The company announced the raise alongside a rebranding—it was previously known as Acclara AI—though as of late April, the transition remained incomplete. Marketing materials live at quillon.ai, but the product itself still answers to app.acclara.ai, a reminder that even well-funded startups sometimes move faster than their domain registrars.
The Audit-Ready Gambit
What Quillon is selling, essentially, is defensibility. The platform combines natural-language search across EDGAR filings, analysis of contracts, and drafting of accounting memos—all anchored by what the company describes as paragraph-level citations to US GAAP and IFRS standards. The pitch: every conclusion the AI reaches can be traced back to a specific regulatory source, making the output legible to the auditors who will inevitably question it.
"We built Quillon to close that gap: a workspace where the AI does the analysis, the accountant drives at every step, and every claim traces back to the exact paragraph in the standards," CEO Nikolay Dakov told Tech.eu in a statement.
It's a narrow ambition compared to some of the broader automation plays flooding the accounting software market, but perhaps a strategic one. Quillon isn't trying to replace accountants—it's trying to make their most tedious, high-stakes work faster and more defensible. The company self-reports SOC 2 compliance and claims not to use customer data for training, two reassurances that matter in an industry wary of both breaches and black-box AI.
More than 500 accountants reportedly use the platform according to Quillon's own claims. That figure, like most startup-reported user counts, lacks independent verification.
Who's Building It
The founding trio brings an unusual mix of credentials. Dakov, the CEO, holds a master's in economics from Oxford and previously worked at Goldman Sachs automating derivatives in the bank's fixed-income division. Ivaylo Stefanov, the CTO, comes from Citibank, where he was a vice president working on large language model infrastructure. Atanas Dobrev, the COO, spent time as a developer for Bulgaria's National Revenue Agency—a background that may prove useful for a company trying to navigate the bureaucratic thickets of global accounting standards.
The startup operates through a US parent that owns a Bulgarian subsidiary, with engineering concentrated in Sofia and operations running out of San Francisco. It's a structure increasingly common among European startups courting American customers and investors, though it can complicate questions of jurisdiction and regulatory oversight.
More recently, Quillon added Alice Yagolnitser, a CPA and former head of technical accounting at Forte, to lead sales and community efforts. Anh Tran, another CPA who has worked on revenue recognition for clients including Atlassian and Palo Alto Networks, joined as a solutions partner. The hires suggest a company trying to build credibility with practitioners who remain, understandably, skeptical of AI in high-stakes financial work.
A Crowded, Well-Funded Field
Quillon's $1.5 million may sound modest, but it arrives in the middle of an investor rush toward AI-powered accounting tools. Rillet pulled in $25 million from Sequoia last May for general ledger automation. Maxima raised $41 million in November for what it describes as an agentic enterprise accounting platform. Even the Big Four are moving—Grant Thornton announced a partnership with Tellen in February, and Thomson Reuters has been expanding its audit ecosystem with a series of AI tie-ups.
What distinguishes Quillon, at least in theory, is its focus on technical accounting rather than broader automation. The company isn't trying to reconcile bank statements or categorize expenses. It's targeting the judgment-heavy work that requires nuanced interpretation of standards—and documentation that can survive regulatory scrutiny.
Whether that narrow focus is a strength or a ceiling remains to be seen. Forbes Bulgaria reported that during due diligence, 42CAP and other investors conducted interviews with accounting experts in the US to validate the product's market fit. That level of caution suggests investors understand the stakes: building AI tools for accountants isn't just a technical problem, it's a trust problem.
What's Next
The fresh capital will fund expansion of both engineering and go-to-market teams, according to coverage in Tech.eu and Forbes Bulgaria. Quillon's roadmap, as described in those outlets, includes evolving from a research tool into a platform capable of producing finalized deliverables and managing entire accounting workflows. Longer-term plans involve expanding into quarterly and annual financial disclosures, the kind of high-volume, high-visibility work that could either prove the platform's value or expose its limitations.
For now, Quillon is betting that accountants—perhaps the most risk-averse users in enterprise software—will trust an AI system if it shows its work. It's a reasonable hypothesis. Whether it's a fundable business at scale is the question every pre-seed round leaves unanswered.
