Born One Year Before Communism Fell. Built in the Age of AI.
Andrew Antos grew up in Brno, Czech Republic - a city better known for producing scientists than startup founders. His mother never got anything lower than an A, all the way through her PhD. His grandfather was a chemist. His family spans four generations of people who understood the world through data, measurement, and proof.
He studied law instead. Not out of rebellion, but out of a calculated bet: technology expertise would make him a different kind of lawyer. His undergraduate thesis at Masaryk University covered the regulation of genetic information - a subject that would puzzle most undergraduates even today.
While interning at a Czech government office working on open data legislation, he spent hours converting government documents between formats, trying to make information useful. That experience planted a seed: a passionate, specific hatred for PDFs. Not a vague frustration - a working theory about what was wrong with the world's information infrastructure.
"Legal contracts run the world and yet only lawyers enjoy reading those."
- Andrew AntosHe went to Squire Patton Boggs in Washington D.C. next, working technology M&A and IT licensing deals. He lasted about a year. The work was fine. The model wasn't. "You're not creating independent value," he observed. "You're selling your time for money." He left for Harvard.
Three Hours Walking Around Cambridge
In 2016, Andrew Antos enrolled in Harvard Law School's LL.M. program and cross-registered for MIT Sloan's New Enterprises course - professor Bill Aulet's legendary class for people who want to build companies, not just study them.
On the first day, he met Nischal Nadhamuni. An MIT computer science graduate specializing in natural language processing and computer vision. In 2016, when most people hadn't heard of transformers, Nischal was already deep in NLP research. The competition for machine learning classes at MIT that year was so intense it had become a minor scandal. Andrew recognized immediately what that signal meant.
They spent three hours walking around Cambridge that night. By the time they stopped, they had a company. Or at least the idea of one.
"We immediately hit it off, spent three hours that night walking around Cambridge and decided to collaborate."
- Andrew Antos, on meeting co-founder Nischal NadhamuniThe initial concept: use AI to read legal contracts, specifically non-disclosure agreements, and automate the review process that junior associates spent enormous hours completing. The logic was clean. The timing was early - maybe too early.
Klarity: From Legal Automation to Enterprise Intelligence
They founded Klarity in 2017. Andrew had the legal domain knowledge and the vision. Nischal had the technical depth to execute it. What they didn't have was a willing customer base.
Law firms, it turned out, were the wrong market. Not because they didn't have a document problem - they had a massive one - but because they didn't want it solved. "Lawyers are the only people who actually enjoy reviewing documents," Antos would later note. "It's their main job, and they don't want to automate it away." Slow purchasing cycles compounded the problem.
From 2017 to 2019, Klarity operated closer to a consulting firm than a SaaS product - solving whatever problems customers brought them, staying alive, accumulating knowledge about where document pain actually lived in organizations. Andrew applied to Y Combinator. Got rejected. Applied again. Got rejected again. Applied a third time. Finally got in, joining the W18 batch.
The pivot came in 2020. Accounting and finance teams at companies like CrowdStrike and DoorDash had the same document problem lawyers had, but without the career incentives to keep it manual. Revenue recognition under ASC 606, contract review for compliance, invoice processing for audit - mountains of documents that needed to be read, compared, validated, and filed. Nobody enjoyed doing it. Everyone wanted it automated.
"Every additional day a startup survives increases the probability of meeting the right market moment, as technology and market preferences continuously shift around you."
- Andrew AntosProduct-market fit arrived like a gear clicking into place. The technology had matured. The problem was urgent. The buyers had budget and authority. Klarity signed major customers and began building toward the category it would define.
Easy, Hard, Three Days
Antos describes Klarity's fundraising arc with the clarity of someone who lived through each phase rather than just reported on them. The seed round was relatively straightforward. Then YC - three attempts. Then Series A in 2022, which was hard: pre-AI hype, a company mid-pivot, asking investors to believe in something the market hadn't yet named.
They closed $18 million. Then, in June 2024, everything changed.
Klarity Funding Timeline
The Series B closed in three days. Seventy million dollars, led by Nat Friedman (former GitHub CEO) and Daniel Gross (former VP of AI at Apple), with Scale Venture Partners, Y Combinator, Tola Capital, Picus Capital, and Invus Capital all joining. Total funding: $90 million.
The contrast between the two rounds - one difficult, one done in days - wasn't just about timing. It was about Klarity having built something real while others were still pitching decks. When AI became a mainstream investment thesis in 2023, Klarity already had the customers, the metrics, and the architecture to back its claims.
"This Series B funding will enable us to hire the best people to scale Klarity to a $100B+ company and deliver exponential value to our customers."
- Andrew Antos, on the Series B closeWhat Klarity Actually Does
The pitch has evolved, but the core insight hasn't: enterprises run on documents, and those documents contain decisions waiting to happen. Contracts tell you revenue. Invoices tell you costs. Purchase orders tell you commitments. Most companies can't read their own documents fast enough to act on what's in them.
Klarity's platform attacks this in two directions. First, intelligent document processing: extracting structured data from contracts, invoices, and purchase orders with 82%+ automation pass-through rates, integrating with CRM systems (Salesforce), ERP platforms (NetSuite), and billing infrastructure. Second, process intelligence: mapping how finance and operations teams actually work, identifying gaps and inefficiencies, and providing what the company calls a digital twin of the organization's processes.
The result for customers: one company reduced a 60-day process to three days. Another eliminated 98% of manual review volume. Finance teams close their books 2-3 days faster. The time savings don't just reduce costs - they change what accountants can do with their time. Instead of processing documents, they analyze them.
In 2025, Klarity launched version 4.0, introducing three new capabilities: Analyst, Advisor, and Coach. The frame shifted from automation vendor to what Antos calls an "always-on AI analyst, advisor, and coach" - closer to an enterprise transformation platform than a document processor. The category: enterprise AI for finance and operations.
The Discipline Behind the Company
Andrew Antos manages time and energy with the methodical attention his mother brought to mathematics. Every meeting starts and ends on time. His calendar is the single source of truth. Workouts - daily, either 6-7am or 7-8pm - are scheduled like board meetings. He uses a treadmill desk for at least five hours every day. He doesn't eat until noon.
Three things matter, in order: family, Klarity, physical well-being. The ranking is a decision, not a declaration.
If your schedule isn't filled with talking to customers and deciding what to build, you're doing something wrong. Everything else is secondary.
Startups move one step forward, two steps back. Being anti-fragile - treating setbacks as expected inputs, not failures - is the most important skill.
The best and most powerful ideas are really simple and straightforward. Complexity is usually a signal that the idea isn't right yet.
He re-discovers annually that networks and relationships matter more than he thought the previous year. The insight compounds, and so do the returns.
On hiring, Antos looks for three things: general aptitude and raw intelligence, strong collaboration and empathy, and curiosity. "There's no substitute for broad intelligence, especially as you're figuring things out." He puts candidates in informal settings - "break bread" - because authentic personalities surface when the stakes feel lower. One reliable red flag: candidates who ask no questions.
"Once you put people in an environment that doesn't feel like an interview, their true colors always come out."
- Andrew AntosHis biggest early mistake, by his own account: hiring too many individual contributors before building a proper management layer. He now emphasizes the executive layer as foundational infrastructure, not a later-stage luxury.
Andrew Antos at the SF Summit Keynote
SF Summit Keynote: Redefining Best-in-Class Accounting Benchmarks with AI (2025)
Klarity 4.0 Launch Keynote - Palo Alto, 2025
The Exponential Organization
The version of Klarity that Andrew Antos is building toward is not an accounting tool. It's an enterprise operating system. The goal: give every company the ability to understand itself in real time - to map its own processes, identify its own inefficiencies, and continuously improve them with AI as the engine.
He calls this the "exponential organization": a company where employees work faster, work better, and spend less time on the 80% of their job that is mechanical. The 80-20 problem he has identified is precise: 20% of enterprise AI adoption is about buying software. 80% is about change management, process redesign, and role restructuring. Most companies stop at the 20%.
The customers who have gone further tell the story better than any metric. Finance teams that spent three days on manual close processes now spend half a day. Revenue operations that required 20,000 manual transactions per year have automated themselves down to hundreds. The time doesn't disappear - it converts to analysis, strategy, and decisions that weren't previously possible.
At 150 employees, with $90M in the bank and a roster of customers that reads like a who's-who of enterprise SaaS, Andrew Antos is running a company built on the same intuition he had while staring at government PDFs in a Czech office in 2015. He was early. He was patient. He was right.