Stealth Startup (Unnamed)

Vik Singh, a successful entrepreneur who previously founded Infer (acquired by Ignite) and now serves as an EIR at Sutter Hill Ventures, teamed up with Danny Ryan, a self-driving car engineer from NIO, to tackle a different problem entirely. While self-driving cars have consumed billions in R&D from tech giants, Singh realized the real breakthrough wasn't building autonomous vehicles—it was taking the vision and deep learning advances that made those cars possible and applying them to a completely different domain: automating the work humans do on computers.

About 18 months before the interview, Singh and Ryan entered R&D mode. Singh started first with the idea, then brought Ryan on as a co-founder. Rather than hiring RPA engineers to write rigid rule-based automation (like UiPath requires), they wanted to build a system that could watch what humans do and learn to replicate those actions using AI. Starting code development roughly a year into the project, they leveraged existing deep learning capabilities developed by major tech companies. The challenge was translating computer vision algorithms from steering wheels and gas pedals to keyboards, mice, and screens—turning pixels into actionable automation that could understand UI elements, read text via improved OCR, and execute mouse and keyboard commands.

In the last six months before the interview, they brought on customers. Singh disclosed they had "tens of customers," many with over $10 billion in capitalization. Their beachhead market was the due diligence and expert research space—an industry where large consulting firms charge $1,000–$2,000 per expert meeting and employ armies of associates to manually find contacts, set up calls, and negotiate incentives. Singh saw this as a massive optimization problem that humans were solving inefficiently. Instead of building the infrastructure and letting customers plug in their own automation (the Zapier/RPA model), they wanted to build the killer applications themselves, starting with automating the expert-finding and meeting-scheduling process.

The core insight resonated: financial firms, private equity, VC firms, M&A advisors, and investment bankers all faced the same expensive, manual, error-prone diligence process. By automating the search and outreach, they could find the right people faster, match them more accurately, negotiate incentives algorithmically, and scale to hundreds of meetings instead of just a few. Singh acknowledged that while Zoom Info solved list-building (finding *who* to target), his platform solved the harder second problem: actually breaking into those accounts and getting them to respond—essentially automating the consultant's job. He noted regulatory constraints in sales/marketing (can't pay someone to talk about your product due to conflict of interest) but found the diligence space had fewer barriers.

After 18 months in stealth, with tens of paying enterprise customers, Singh and Ryan have bootstrapped the company using proceeds from their previous acquisitions rather than raising traditional venture capital. They've decided not to disclose exact equity splits or funding amounts. When asked about raising, Singh said it "feels realistic" based on current momentum, but they're not in a rush—they're profitable or cash-flow positive and still exploring whether to go enterprise-focused or open the platform to a broader market. The decision on that positioning will dictate their fundraising strategy.