Jeff
Seibert

In the middle of running Crashlytics - one of the fastest-growing mobile developer tools ever built - Jeff Seibert kept bumping into the same absurd problem. He ran a software company that knew, to the millisecond, when any of the 500 million apps using his SDK crashed. But he had to wait two to three weeks after each month ended to learn whether his own company was profitable.

That gap - between what technology could do and what finance actually delivered - became the founding thesis for Digits. But the story starts much earlier, in a Baltimore bedroom in 1998, with a book called Mac Programming for Dummies and a 13-year-old who had already decided that "ship it" was a more interesting philosophy than "wait and see."

Jeff Seibert released Histogram - a specialized graphing application for Mac OS - at age 13. Not as a school project. As shareware. By high school he had shipped EVONE, a graphical editor for the cult game Escape Velocity. These weren't toys. They were products, put into the world to be used by strangers. That instinct - build, release, iterate - has defined every chapter that followed.

At Stanford, he didn't just study computer science. He ran the Entrepreneurial Thought Leaders seminar series as co-coordinator, earning a front-row seat to the founders who would shape the next decade of tech. He was a Mayfield Fellow in 2007 - the same year he co-founded his first company, Increo, a document collaboration startup, before he'd even graduated.

Increo's fundraising story is almost too good. Two minutes into a pitch meeting with Tim Draper, the legendary venture capitalist cut to the chase: "Okay, how much are you raising?" No deck required. They raised $500,000 on the spot. The company was ultimately acquired by Box in 2009 - with roughly 30 days of cash left in the bank - and its document rendering technology quietly powered Box's platform for five years.

The premise of Digits is not subtle. Seibert watched Twitter's 100-person finance team struggle to answer a basic budget question in real time - and realized that if the most sophisticated companies in the world couldn't do this, the problem was structural, not just a matter of hiring better accountants.

Most accounting software was built for accountants. Digits was built for founders - people who need to know what's happening with their money right now, not when the books close at month-end. The product converts millions of raw transactions into what Seibert calls a "living model" of the business: interactive, real-time, and increasingly driven by AI that has been trained on $300 billion in small-business transaction volume.

The technical architecture is deliberately unusual. Rather than exposing a raw LLM to financial data, Digits built a three-layer system: a proprietary financial modeling engine with AES-256 encryption, custom-trained deep-learning models for accounting-specific tasks, and public LLM APIs that receive only anonymized customer identifiers. This structure lets the product use the best of generative AI without surrendering the security guarantees that financial data requires.

The comparison Seibert makes is to mobile. He argues AI's impact on every industry will exceed what mobile technology achieved - a claim he makes with some credibility, given that he literally built infrastructure that runs on more mobile devices than almost any software in existence.

The product roadmap has moved fast. In a single month in 2024, the Digits team shipped nine new features - including v1.1 of the iOS app and a new Model Context Protocol (MCP) server that connects the Digits financial engine to AI assistants. In February 2025, they added AI-powered invoice processing. The company's customers have noted the shipping velocity publicly.

Seibert co-founded Digits with Wayne Chang - the same person he co-founded Crashlytics with. Two exits, two companies, same founding partner. At some point, pattern recognition becomes design choice.