Two years ago today, I made one of the most difficult decisions of my life: starting linehaul.ai. While hundreds of thousands have started their own Freight Brokerages, this was different. We were two and a half years into what was (and still is) statistically the worst freight recession since deregulation, and I was building a company on the predicate of automation and a very unproven technology—Generative AI.
I was coming off a VP of Sales role at my previous company and always knew that one day I would go out on my own. In my head, the perfect year to launch my own 3PL was 2025 or 2026. But after two weeks of taking interviews at logistics and tech companies, weighing my options against the risk to the financial health I had worked so hard for, it hit me: there’s no perfect time to start a business. Just like there’s no perfect time to buy your first house or have your first child (I’ll have to take your word on the latter).
The very next day, I drove downtown to the Illinois Secretary of State and registeredLINEHAUL TRUCKING LLC, d/b/alinehaul.ai
I was acutely aware of the high failure rate for startups. To scare myself a little bit more, I did a deep dive into the statistics before spending the next fourteen days creating budgets, calling vendors, and absorbing every piece of information I could from intelligent people I trust – thank you Jake Elperin, Robert Rajfer, Marco Vargas-Avila, and many more.
My plan was to go at it alone for the first month. The only prospect worse than failing is convincing others to join you and then costing them their livelihoods. Fortunately, my plan was short-lived. When one of my former Account Executives, Anya Crane, asked to join and I told her “absolutely not,” she took it upon herself to resign the next day and message me, “Now you have to hire me.” That was our “Jerry Maguire” moment (apologies to any Gen Z readers for dating myself).
Initially, I figured we’d use an off-the-shelf TMS and I would retrain and fine-tune the LLaMA foundation model with our own data and publicly available logistics data. Well, after just two months of using that TMS, I decided we needed to build our own. The challenge? I hadn’t written a line of code since 2004, and nobody was building a modern TMS in C++. I won’t bore you with the technical details, but we built our own TMS in just forty-five days. Since then, we've continued to build several internal platforms that empower our team to move our customers' freight to the best of our ability. We’re an automation company, but we are a freight brokerage first, and our team works around the clock because we take every single load personally.
If you follow our social media, you’ll see we’re not ones for overt self-promotion(perhaps to a fault). This post isn’t a method to acquire new customers or attract investors (we’re proudly bootstrapped 🙃). It’s a thank you. A thank you to our team, friends, and colleagues who have helped me along the way – shoutout to Paul Lang, William Halloran, Chris Ceausu, Ken Miller, Garrett Allen, Nicholas Sorrell, Howard Abrams, Krystian Gebis, Paul-Bernard Jaroslawski & more that I'm forgetting.
Most importantly, thank you to our CUSTOMERS and CARRIERS. I am incredibly grateful that every day, I get to do business with honest, hard-working, and well-intentioned people.
It hurts to see so many logistics companies go under over the last two years, but I’m thankful we’ve been able to make a difference in our corner of the industry. I pray for a brighter future ahead, and mark my words, we will be giving back soon enough when we open-source the bulk of the software we’ve developed.