Y Combinator F25 · San Francisco
Zarna
The deal team that never sleeps.
YC F25
AI Agents
Private Equity
Fintech · B2B
The deal team that never sleeps.
Every private equity firm runs on a small army of junior associates who tear down CIMs, build LBO models, update CRM records at midnight, and draft IC memos before 7am. It is expensive, time-consuming, and - let's be honest - a significant chunk of that labor is pattern-matching work that a well-trained agent could handle.
Zarna builds those agents. Not generic AI tools bolted onto existing software, but autonomous associates wired directly into a firm's data, document history, and institutional memory. They learn how your firm does deals. They work off your templates, your mandate, your decades of proprietary pattern recognition.
And unlike the junior analyst you just hired, they do not call in sick on a Friday afternoon when the banker just dropped a hot CIM.
"The future belongs to firms whose intelligence compounds faster than their capital."- Zarna
Zarna ships four agents, each covering a distinct part of the deal workflow. They operate semi-autonomously, triggered by real events - a new email, a dropped document, a flagged deal - and deliver finished outputs directly to the tools your team already uses.
Captures, structures, and logs deal information across Outlook, Teams, and Slack. Parses emails, documents, and meeting notes. Deduplicates contacts and companies so your CRM stays clean without anyone touching it.
Generates firm-formatted documents: screening memos, banker call briefs, weekly deal flow reports. Uses your firm's own templates and voice - not a generic output that gets redlined into oblivion.
Scores inbound deals against your mandate. Pattern-matches against your historical deal history. Flags risks and opportunities before the first associate has opened the CIM.
Maps your firm's network to surface warm introductions and domain experts relevant to an active deal. The rolodex, automated.
Before Zarna was a company, it was a problem. Four UC Berkeley alumni - Rishabh Dhariwal, Hrishi Joshi, Rakesh Mehta, and Vivan Agrawal - were embedded inside AEA Investors, a $20 billion private equity firm, as forward-deployed engineers. Their job: lead AI enablement initiatives. Automate core deal workflows end-to-end. Build agents across sourcing, diligence, and portfolio management.
They saw exactly where time, judgment, and firm-specific knowledge were getting bottled up. They saw which workflows were eating analyst bandwidth and which ones could be handed to a well-designed agent. And then they left - and built it for everyone.
The founding team brings experience from Deutsche Bank, Google, Amazon, SAP, Nutanix, Wipro, and EY. Rishabh's advisory work alone covers Pepsi, Rajasthan Royals (yes, the IPL cricket team), Roche, and LATAM Airlines. This is a team that has seen how deals actually move inside large institutions - and built software accordingly.
Economics, UC Berkeley. Led AI enablement at AEA Investors. Prior: Deutsche Bank, advisory work across Pepsi, Roche, LATAM Airlines, Rajasthan Royals. Reads geopolitics and cricket between board decks.
UC Berkeley. AI engineering background with experience across enterprise software. Built agent infrastructure at AEA Investors before co-founding Zarna.
UC Berkeley. Systems-level engineering background. Prior experience spans large enterprise environments before embedding in private equity workflows.
UC Berkeley. Leads operations and go-to-market. Brings a blend of engineering and business context from the founding team's time at AEA Investors.
"Investors spend less time buried in decks and spreadsheets, and more time on judgment, relationships, and winning deals."- Zarna's founding thesis
Zarna is not trying to sell to every financial services firm. They have mapped a specific corner of the market: private capital teams where lean deal teams are doing disproportionate work, where the ratio of hours spent on analysis to hours spent on actual judgment calls is badly skewed.
Their addressable market: roughly 3,000 firms, at approximately $75K per year in licensing. That is a $225M TAM before international expansion. Focused, defensible, and deep enough to build a serious business.
Current private beta includes multi-billion AUM PE firms and large M&A advisory companies. They are not splashing the customer list - that is how sophisticated finance buyers like it.
The founding team embeds at AEA Investors - a $20B PE firm - as forward-deployed engineers. They build AI agents automating deal sourcing, diligence, and portfolio management workflows from the inside.
Rishabh, Hrishi, Rakesh, and Vivan leave AEA and incorporate Zarna in San Francisco to bring what they built for one firm to the entire private capital industry.
Accepted into Y Combinator's Fall 2025 batch (F25). Raises $500K in seed funding. Begins private beta with multi-billion AUM PE firms and large M&A advisory companies.
Ships four-agent platform - Scribe, Memo, Analyst, Experts - covering the full deal lifecycle. Achieves SOC 2 Type II certification. Launches 30+ integrations including DealCloud, Affinity, CapIQ, PitchBook.
Continuing to expand private beta. Building toward standardized onboarding using pilot case studies and partner channel relationships.
AES-256 encryption at rest. TLS 1.3 in transit. Single-tenant infrastructure with strict data isolation - your deal data never lives next to another firm's. Role-based access control with SSO. Data is never used for model training. Penetration tested with regular audits. These are not checkbox compliance features; they are table stakes for selling to institutional finance.
The founders were not just customers of the problem - they were embedded inside a $20B PE firm, building the very systems that would eventually become Zarna. Very few founders can say they field-tested their product before the company existed.
Rishabh Dhariwal's advisory portfolio before Zarna includes Pepsi, Roche, Rajasthan Royals (IPL cricket), and LATAM Airlines. A healthcare company, a cricketing empire, a soft drink giant, and a South American airline - not a typical pre-startup resume.
Zarna offers a genuine walk-away clause: if ROI targets are not met within three months, you leave. Enterprise SaaS does not usually put that in writing. It tells you something about how confident the team is in the actual product.
The name: "zarna" in Gujarati and Hindi means a spring, a natural fountain - a continuous source. Intentional or not, the metaphor fits: not a one-off tool, but a persistent, always-on source of intelligence flowing through the firm.
All four co-founders attended UC Berkeley, studying computer science, electrical engineering, and economics. The intersection of those disciplines - systems thinking, quantitative modeling, market understanding - shows up directly in how Zarna's agents are designed.
The team has published thinking on agent design philosophy: when the top 5 models on SWE-bench differ by less than 1 point, model selection is not the question. Framework and harness design is. That kind of focus shows up in the product's architecture.