Built in 2013. Rebuilt for 2026.

In 2013, I built a company called ReplyWire.

The pitch was simple. Every business has contact forms, newsletter signups, live chat, and a dozen other ways people reach out. And every one of those channels was a leak. Messages got misfiled. Nobody knew if a reply went out. Teams couldn't scale their response without hiring more humans. Contact-to-customer conversion was quietly bleeding out in the gap between "someone reached out" and "someone actually followed up."

I shelved ReplyWire before it shipped. The tech wasn't ready, the market wasn't ready, and frankly, neither was I.

Thirteen years later, I'm running three companies with lead funnels spread across six platforms, and I just realized I'm losing real leads because I can't see them in one place. The exact problem I tried to solve in 2013 is the problem I have today. I've talked to enough founders and operators to know I'm not alone.

Here's what's different now. AI can do the job we needed humans to do in 2013. Triage. Classification. Draft replies. Routing. Pattern recognition across thousands of messages. The "auto-responder" we pitched in 2013 was a dumb template. The auto-responder in 2026 is a research assistant that reads every message, understands what it is, drafts the right response, and tees it up for you to approve or send.

That's what ReplyWire is in 2026. One inbox for every way your customers reach you, with an AI layer that keeps you from dropping anyone. I'm building it for my own companies first. If it works, I'll open it up.