Matthew Tillman and the cost of looking the other way

Matthew Tillman wanted to send a friend’s truck to Argentina. The ambition was to bring it back north along the Pan-American Highway, all the way to Alaska. First, though, somebody had to get the truck onto a ship. An adventure that might reasonably have begun with a map instead began with freight forwarders and an inbox.

As he asked for prices, Tillman found himself conducting a reverse auction by email. Booking space exposed a market whose arrangements seemed unnecessarily difficult for the buyer to understand. He had come looking for transport. He stayed interested in the transaction. It is a useful distinction when the eventual result is a software company.

A truck, a ship, and an inconvenient question

The family connection makes the episode more interesting. Tillman is the son and grandson of truckers, the first in three generations to make his living elsewhere. The industry was part of his family’s experience before it became his market. His career would bring him back to freight through the keyboard rather than the cab.

In 2014, he founded Haven to help commodity businesses automate logistics and collaborate with their partners. Its ambition stretched across the work of shipping cargo, rather than stopping at a single booking. By 2018, Haven had offices in San Francisco, Singapore, and Switzerland. The business depended on understanding how people in different places coordinated the same shipment.

That setting matters. A cargo movement involves decisions, documents, counterparties, and deadlines that have to agree with one another. Software enters a world that already has habits and obligations. For Tillman, logistics offered something tangible: goods had to move, and mistakes had consequences outside a screen. His proposed road trip had become an education in what happened before anything reached the road.

Learning which side of the transaction to serve

He had already spent years making software. His early career included development work at Communicative Design Technologies and Starbelly.com, followed by Newedge. Product roles took him to Adchemy in 2007 and SET Media in 2010. At SET, he became vice president of products. Finance and advertising supplied an apprenticeship in systems where information, pricing, and commercial incentives meet.

The progression from developer to product executive helps explain the continuity across otherwise different industries. Trading, advertising, and cargo do not share much scenery. Each requires a business to make decisions from records that arrive through other people and systems. Tillman’s work repeatedly places software between that incoming information and the decision somebody has to make.

Haven also gave him an education in how to build the organization around the product. He advocated remote teams and outsourcing early in a startup’s life, keeping spending directed toward the work that made the company useful. A global trade business, he argued, benefited from people with different languages, backgrounds, and proximity to its market.

There was a financial logic, too. An expensive local office and a locally concentrated team could consume capital that might otherwise improve the product. Tillman’s argument asked founders to decide which capabilities they needed to own and where outside help made sense. The office postcode was a poor substitute for that decision. A business could look the part while spending too much to play it.

He was candid about hiring mistakes at Haven. Industry knowledge did not automatically translate into an ability to sell software. The qualities he emphasized were curiosity, intelligence, and excitement about the vision. His instruction was brisk enough to fit on a recruiting form.

“Never ever hire for a rolodex.”

Matthew Tillman, on building Haven

The invoice is where the argument lands

With Parker Moore, Tillman co-founded OpenEnvoy in 2020. The company put a specific commercial problem at the center of its work: determining whether an invoice agrees with the documents behind it before the customer pays. The bill becomes a meeting point for the contract, the delivery record, the agreed rate, and the money about to leave the business.

The early financing followed quickly. OpenEnvoy announced a $6.5 million seed round led by Riot Ventures in April 2021. RRE Ventures, Gokul Rajaram, Naval Ravikant, and TNT Ventures were among the participants. Tillman’s message at the time tied accurate billing to understanding purchases and maintaining supplier relationships. The opportunity extended beyond finding a stray duplicate.

A $15 million Series A led by RRE Ventures followed in May 2023, with Coelius Capital, Hack VC, Riot Ventures, and Uncorrelated Ventures participating. The announcement emphasized complicated, high-volume invoices, including the variable costs found in logistics and manufacturing. Across the two disclosed rounds, the company had raised $21.5 million. The money financed a business designed to scrutinize where other businesses’ money went.

In explaining OpenEnvoy, Tillman has returned to the burden of checking an invoice against its supporting evidence. Capturing a number from a document leaves several questions unanswered. Was that the agreed price? Does the delivery information support the charge? Does the record belong in the accounting system as approved? His case for machine learning begins with the effort required to answer those questions consistently.

He has described the alternative as a loss of control: elaborate approval chains pass responsibility outward while finance remains responsible for the result. His stated ambition is to perform the audit before the invoice reaches the enterprise resource planning system. That moves the opportunity to challenge a charge to a point when the buyer still holds the money.

A chair arrives late. The bill arrives anyway.

Tillman’s supply-chain commentary occasionally comes with a domestic prop. Discussing holiday logistics in late 2023, he described a desk chair he had ordered for November delivery that was running three months behind because of stock problems. The executive explaining the system was also the customer waiting for somewhere to sit.

His wider point concerned the pressure that uncertainty puts on purchasing. Under time pressure, evidence can remain buried in email, arrangements become improvised, and invoices get paid before anyone has checked them closely. The cost of moving the goods and the cost of failing to inspect the bill then arrive together.

He draws a practical distinction between preventative auditing and recovering money later. A correction before payment can become an ordinary conversation with a supplier. A correction after payment can require a clawback and a wait. The commercial relationship remains important in both cases. Timing changes how much awkwardness the buyer has purchased along with the service.

Integration is part of that argument. Tillman has said the company intends to work with existing systems and maintain the integrations it builds. For a buyer, a useful verification tool must fit into an operation that already has accounting software, suppliers, and established routines. Otherwise, the proposed efficiency begins with a lengthy new assignment.

The numbers have to survive the sales pitch

By September 2025, OpenEnvoy was describing work across more of the finance operation. In its account of customer Schreiber Foods, the company reported 90 percent accounts payable automation within a month and cash application moving from three or four days to the same day. These were company-reported customer results, attached to a particular deployment. They give the argument a concrete setting without promising every customer the same outcome.

Building that business has also meant building a wider team. The same announcement introduced Liz Kaiser as vice president of marketing and Theresa Contario as general manager of agency and media sales, while promoting Adam Mills to lead customer success and operations. The operating problem had expanded beyond what two founders could explain or deliver themselves.

Tillman has a parallel role as a founder of TNT Ventures, which focuses on B2B software investing. On his public profile, he says that even without financial constraints he would be starting companies and running a seed-stage fund. It is a fairly revealing answer to a question that might have produced a beach. His imagined alternative working life resembles his actual one.

A hostage, a novel, and the next question

Other public details suggest someone comfortable taking a less solemn view of his own biography. His personal site points visitors toward a cameo as a hostage in the video game Xeno Crisis. It also lists Sea Shepherd, the International Anti-Poaching Foundation, and the arts organization Root Division. His LinkedIn records a stint on Root Division’s board from 2013 to 2015.

His reading preferences pair Andy Grove’s Only the Paranoid Survive with Gore Vidal’s Julian. He rates the latter as useful for business, despite its usual shelf. Beside a career spent simplifying transactions, there is room for a historical novel, conservation, and the possibility of needing rescue in a video game.

The more recent work returns to the practical question of what finance teams can do with checked information. In May 2026, OpenEnvoy joined the Sage Intacct Marketplace. Tillman described a verification layer positioned before an invoice becomes an entry in the accounting system, available within software that finance teams already use.

In June, the company launched Ask OE, a natural-language interface with matching, document, and analytics agents. Its premise puts verified records underneath the conversation. Finance professionals bring their context and questions; the software gives them ways to investigate exceptions, correct extracted information, and work with financial data. The new interface extends a familiar concern: an answer is only useful if the information underneath it deserves the user’s trust.

The truck-to-Argentina episode remains an apt beginning because it started with a customer trying to accomplish something ordinary. Tillman’s professional route has become more technical, and the documents have multiplied. The question still fits on a bill: does what we are being asked to pay match what we agreed to buy?