Xiaohan Shen wrote 1,000 cold emails by hand. Then he built the machine he wished he'd had.

There is a particular kind of tired that comes from writing a good cold email. Not the template kind - the real kind, where you read a stranger's company page, figure out what they might actually need, and try to say it in three sentences that don't sound like every other pitch in their inbox. Xiaohan Shen knows that tired well. Before Coldreach was a company, before Y Combinator, before the phrase "custom intent signals" appeared on a pitch deck, there was just Shen and his co-founder at a desk in Madison, writing those emails one at a time. More than a thousand of them.

That number is the origin of everything. It is easy to describe Coldreach as an AI sales-prospecting platform and move on. It is more honest to say it started as a complaint - a specific, lived frustration that the best way to reach the right customer was also the slowest thing in the building.

01 · The classroom partnershipTwo students who kept building things

Shen studied computer science and data science at the University of Wisconsin-Madison, graduating in 2023. Somewhere in the lecture halls he met Hankel Bao, a fellow student with a coder's biography that reads like a dare - Bao reportedly started programming at nine and built an operating system in middle school. The two of them clicked over the thing that clicks founders together: an inability to leave an interesting problem alone.

Their first company was EasyJobs, an AI-powered assistant for job seekers. It found relevant openings and auto-filled applications, shaving hours off the miserable ritual of job hunting. It worked. EasyJobs climbed to roughly 20,000 monthly active users in about a year and a half - the kind of traction most student projects never sniff.

But a growing product still needs customers, and customers needed to be reached. So Shen did the unglamorous work himself. He wrote the outreach. He learned, in his bones, which emails got replies and which got ignored. He became - his words - an "outbound nerd." It is a self-description he offers without irony, the way some people admit to loving spreadsheets.

02 · The pivotLeaving a good thing to chase the real thing

Here is the founder move that is easy to miss. Most people with a 20,000-user product would double down on it. Shen and Bao noticed something else instead: the painful part of their own business - the manual, personalized outreach - was a problem every other B2B company had too. They had accidentally become experts in something bigger than job search.

So they pivoted. The new company, Coldreach, would take the thing they hated doing by hand and make it scalable for everyone. It got into Y Combinator's Winter 2023 batch - the same accelerator that launched Airbnb, DoorDash, and Instacart - and the pair moved the operation to San Francisco. Shen took the CEO seat; Bao runs engineering and the natural-language work as CTO. Same partnership from the classroom, bigger stage.

03 · The ideaTiming is the product

The insight that makes Coldreach more than "an AI that writes emails" is quieter than it sounds. The problem with cold outreach was never only the writing - it was the guessing. Who, out of thousands of possible companies, actually needs your product this week? Spray-and-pray answers that with volume. Coldreach answers it with signals.

The platform monitors public signals customized to what a given team actually sells, then flags the moment a company becomes a good fit. A firm hiring its first accountant is a lead for accounting software. A team posting for its tenth sales rep is a lead for sales tooling. Reach out at that moment, with a message that references the reason, and the math changes. Coldreach's own claim is that this kind of targeted outreach books meetings at several times the rate of random campaigns.

04 · The habitDo it by hand first

If there is a lesson worth stealing from Shen's path, it is this: he built the software only after doing the manual version until it hurt. The thousand emails were not wasted effort or a story he tells for color. They were the research. Every reply and every silence taught him what a good signal looked like, what a believable opening line sounded like, and where the whole process broke down. By the time he wrote a product spec, he had already been the product.

“Watch the signals. Reach the right person. At the right time.”

That grounding shows up in how he talks about the work. He is less interested in AI as a spectacle and more in whether a sales rep books a meeting they otherwise would have missed. Across two startups he has built outbound engines from scratch, and along the way he has helped more than 50 sales teams put a consistent motion in place. The trendline of his career is not a jump from idea to idea - it is one obsession, followed patiently.

05 · What's nextMaking the personal touch scale

The tension at the center of Coldreach is also the interesting bet: can a machine keep cold email personal? Automation usually trades warmth for volume. Shen's whole premise is that you can have both - if the machine does the research a good rep would do, instead of just blasting a template with a first name swapped in. Whether that holds at scale is the question the company exists to answer.

From a high school in Beijing to a lecture hall in Madison to a small team in San Francisco, the throughline has stayed the same. Reach the right person, at the right moment, with something worth reading. Shen just decided to stop doing it one email at a time.