Expedock
reads the freight.

The night shift never closes

Somewhere in Manila it is two in the morning, and an Expedock operator is staring at a commercial invoice from a Vietnamese factory addressed to a Long Beach 3PL. The numbers are misaligned. The HS codes are partial. Last week the same shipper used a different template. The AI has already taken its best swing; now a human is grading the AI's homework, and a freight forwarder in San Francisco is asleep, blissfully unaware that the bill of lading on his Monday calendar is already done.

This is the unglamorous engine room of global trade. Expedock is the company that runs it. The pitch sounds small if you say it fast - "we read freight documents" - and enormous if you say it slowly: every container that crosses an ocean drags behind it a paper trail older than the container itself, and somebody has to convert that paper into rows in a database before anything moves.

A $7 trillion industry that runs on PDFs

Global freight forwarding is, by some counts, a seven-trillion-dollar industry. It is also, by every count, an industry that still emails PDF attachments back and forth like it is 2003. Commercial invoices arrive as scans. Packing lists arrive as Excel sheets that have been turned into PDFs that have been emailed as JPGs. Bills of lading come from carriers in formats designed by people who have never met each other.

The freight forwarder's job, in practice, is to take these documents and re-key them into a transport management system - CargoWise, Magaya, Descartes - so that an invoice can be issued, a shipment can be tracked, and an importer can get their cargo. The job is not difficult. It is just relentless, and there are thousands of documents per day per office, and every minute spent typing is a minute not spent winning new lanes.

For decades the answer to this problem has been "hire more people." For the last few years it has been "hire more people in a cheaper time zone." Expedock's bet, made early and stubbornly, is that the right answer is something else entirely: hire fewer people, train an AI to do the boring 80 percent, and put the remaining humans on the parts where boring would be expensive.

Two seatmates and a kid who grew up around forwarders

The Expedock origin story has the slightly suspicious neatness of a story that has been told a few times. King Alandy Dy and Rui Aguiar met as seatmates in a computer science class at Stanford. King had grown up running parts of his family's import-export business across Southeast Asia, China, and the Philippines, starting at sixteen, which is the sort of biographical detail that becomes useful when you decide to start a freight company. Rui had the AI chops. They added Jeff Tan - who, conveniently for the plot, comes from a family of freight forwarders - and Jig Young, the product mind.

Four founders is usually one too many. In Expedock's case it has been arguably one too few, because the company has always insisted on doing two things at once: a software product, and an offshore operation. The first without the second turns into a demo. The second without the first turns into a call center. Together they turn into a moat - or at least the company believes so, and so do its investors.

What it actually does, in plain English

Strip away the deck slides and Expedock is, essentially, a pipe. On one end: any document a freight forwarder receives. On the other end: clean structured data, sitting inside the customer's TMS, ready to be invoiced or tracked. In between: a stack of machine learning models trained on millions of freight documents, an interface for human operators to review whatever the model is unsure about, and a roster of those operators working in shifts to make the throughput round-the-clock.

Captions tend to undersell this kind of thing. The honest one is: it's a lot of unglamorous work, done quickly.

Numbers, partnerships, and one specific decimal place

Customer logos read like a who's-who of mid-market and global forwarding - the kind of names that move actual freight rather than dispatch press releases. The partnerships do most of the talking. CargoWise is the operating system of the freight forwarding industry, and Expedock has built one of the tighter integrations into it. Magaya and Descartes round out the rest of the market. Shipthis, the newer entrant, is a hedge against where the industry might go next.

Talent and AI, without the gatekeeping

Ask the founders what they are building and the answer is usually some variation of: a workforce-augmentation platform that pairs elite offshore talent with cutting-edge AI, so that any business - not just the Fortune 500 - can compete globally. It is the kind of mission statement that survives translation into a pitch deck and a town hall in roughly equal measure.

The interesting part, if you squint, is the implied politics. Most AI companies tell you their tech will replace people. Expedock tells you the AI is there to make a specific group of people - well-paid offshore operators in places like the Philippines - dramatically more productive, and therefore more competitive in a global market that has spent a century telling them to be cheaper instead. Whether you call that a mission or a market opportunity probably depends on which seat you are in.

Why this is a bet worth watching

Freight forwarding will not be disrupted by a single dramatic product launch. It will be eroded, line item by line item, until the people who used to do the data entry are doing exception handling, and the people who used to do the exception handling are doing customer strategy, and somebody, somewhere, is finally answering the email that has been sitting unread since Tuesday. Expedock is one of a handful of companies actually doing this work, and unusually, it is one of the few that has decided to own both the software and the operators who use it.

The risk is obvious. Doing two businesses at once is harder than doing one. Margins on services are lower than margins on software. Customers who buy services sometimes mistake their vendor for staff, and customers who buy software sometimes mistake their vendor for an oracle. Expedock has to walk both lines without falling off either.

The upside, if it works, is that Expedock becomes the boring, reliable infrastructure of an industry that desperately needs boring and reliable. Not the most romantic outcome. But freight, ultimately, is not a romantic business.

Back to the inbox at two in the morning

The operator in Manila has finished with the Vietnamese invoice. The AI got the line items right. The HS code was a 50/50 call and she made it. The bill of lading on that San Francisco forwarder's Monday morning is done before he has poured coffee. He will not notice, which is the point. Expedock is, on its best days, invisible.

That is the whole pitch. A trillion-dollar industry, a stack of paperwork older than the steam engine, four founders with a thesis, a hundred and fifty employees, and one very specific decimal place. The freight keeps moving. Someone, somewhere, is reading every document. It just might not be you anymore.

Where to look next

• expedock.comWebsite
• LinkedIn · /company/expedockLinkedIn
• @ExpedockAITwitter / X
• facebook.com/ExpedockFacebook
• Crunchbase profileFunding data
• TechCrunch: Series A coveragePress
• Insight Partners announcementInvestor note
• Founder interview · King Alandy Dy (YouTube)Interview
• Motion Ventures Founders' SpotlightProfile