Jordan Zamir knew what a contract meant. He had been a corporate lawyer, working on mergers, acquisitions and the patient business of negotiating terms. He understood the machinery behind a transaction. Then he became a founder and discovered how much machinery still required a person to stand beside it, turning the handle.
The distinction bothered him enough to become a company. Zamir is the co-founder and CEO of Turnstile, a San Francisco software business concerned with the distance between agreeing to a sale and collecting the right amount of money. It is a peculiar distance: sometimes only a few clicks, sometimes an archaeological expedition through a company's files.
His own explanation begins with an admission. Familiarity with contracts and revenue operations had given him a head start. It had not supplied a working daily process. Understanding a billing schedule and keeping a business running on one are separate accomplishments. The second involves tomorrow, the month after tomorrow, and the exception somebody agreed to last Tuesday.
“But seeing it and owning it are different things.”
Jordan Zamir, on becoming a founder
An engineer's education, a lawyer's seat
Zamir's education crossed disciplines before his career did. He earned a bachelor's degree in electrical engineering and computer science at the University of California, Berkeley, graduating in 1998 with honors. His profile also lists Tau Beta Pi, the engineering honor society. A law degree from UCLA followed in 2001.
Those qualifications put code and contracts unusually close together on the same résumé. His professional work included corporate transactional law at firms including Jones Day, followed by senior roles at Minted, Second Measure and Scratch. At Minted, he was general counsel. At Second Measure, his responsibilities included legal affairs and people operations.
It is tempting to make that sequence look inevitable. His account is more useful than that. A lawyer can see the consequences of a deal without being the person who must carry every detail into the operating system. Becoming responsible for execution revealed work that subject knowledge alone had not removed. He describes the learning curve as real, even with considerable prior exposure.
He had another unusual advantage: his wife had spent years specializing in revenue recognition at a Big Four firm. By his own telling, some of the relevant conversations happened at the dinner table. That is a fairly specialized household advantage. It still did not spare him the task of turning knowledge into a process that ran without a finance team.
The spreadsheet earns a promotion
Second Measure gave the problem a concrete setting. The transaction-data business was acquired by Bloomberg in 2020. Zamir's later account of operating there includes repeated reconciliation work and spreadsheets that had acquired responsibilities far beyond their modest beginnings. Before a board meeting, the team had to work to establish confidence in its own revenue numbers.
A spreadsheet is an accommodating colleague. It accepts an improvised column, a special discount and an explanation in a cell comment. Its willingness makes it useful. Its willingness also helps it accumulate jobs. Eventually the convenient record becomes the place everyone depends on, while each department keeps a slightly different account of the same customer.
In his launch announcement, Zamir described a signed agreement dispersing into separate interpretations. Sales needed one set of details, customer success another, finance another. The contract survived as a document; the information inside it had to be extracted repeatedly. The same sale could therefore create several small clerical projects, all with an opportunity to disagree.
Turnstile's starting proposition was to preserve the commercial terms as usable data. A negotiated agreement should carry its meaning into billing, subscription management and reporting. That proposition reflects the part of Zamir's career spent living with the consequences of a transaction. A signature settles the negotiation. It also starts an operational obligation.
- 01Agree the termsPrice, usage, timing, amendments
- 02Keep one commercial recordThe promise remains usable data
- 03Carry it throughSubscriptions → invoices → reporting
A company built for the person doing three jobs
In February 2026, Turnstile publicly launched and announced $29 million in funding from First Round Capital, OMERS Ventures, Illuminate Financial and angel investors. Illuminate and OMERS co-led its Series A. The money gave an unglamorous administrative problem a conspicuously large number attached to it.
The launch pitch addressed founders and lean teams that negotiated real business contracts but lacked the people and time to administer a complicated enterprise installation. Zamir presented Turnstile as a self-service route into quote-to-cash software. The ambition was practical: let a team support a customer's negotiated terms without repeatedly rebuilding those terms downstream.
He was working with people familiar with the earlier operating environment. Michael Babineau, a co-founder of Second Measure, also co-founded Turnstile. Illuminate's investment account highlighted Zamir and Lillian Chou's experience at Second Measure. It emphasized product quality and customer feedback, the kind of qualities that become conspicuous when the output of your software is somebody else's invoice.
There is a useful restraint in that choice of problem. Sending a bill is ordinary. Sending the agreed bill after a contract changes halfway through its life takes care. The company's wager is that businesses will pay to stop repeating that care by hand. Zamir has chosen to build around a job whose success often passes without comment.
The number should survive the meeting
His 2026 essays show that the concern extends beyond invoices. He objects to revenue figures that require elaborate verbal assistance. In June, he described a framework he calls “simple-complex-simple”: begin with a straightforward question, do the difficult normalization behind the scenes, then deliver an answer the organization can use.
Annual recurring revenue is his recurring example. Ramp periods, different contract lengths and changing products all complicate the calculation. He argues for resolving those questions consistently within the process. Otherwise, the meeting intended to decide what to do next becomes a seminar on what the previous slide meant.
This is a revealing preference in an executive who once negotiated documents for a living. He wants the difficult work done, but he resists making every listener experience all of it. His concern is decision-making: sales and finance need numbers they can use together when discussing hiring, customers and growth.
“simple-complex-simple”
Zamir's framework for a useful business metric
In July, he returned to the issue from the perspective of AI businesses and usage-based revenue. Consumption can fluctuate; different ways of annualizing it can produce different impressions. His argument allows for consistent analytical definitions while demanding traceable underlying data. A number should connect back to customer activity, rather than depend on someone's latest spreadsheet interpretation.
His emphasis falls on the moment when another party examines the business closely. An audit, an investment process or an acquisition brings fresh eyes to records that colleagues have learned to navigate by habit. The earlier lawyer's world returns here. A claim about revenue needs an account of how it was earned.
Some things are worth buying
Zamir also has a pointed view of the enthusiasm for building internal software with AI. In an April essay, he distinguished a convenient internal prototype from a financial system that must handle actual customer money. A useful experiment can tolerate rough edges. An incorrect bill arrives in somebody else's inbox.
He frames the decision around risk and the use of engineering time. Maintaining ordinary business infrastructure can consume effort a company might otherwise spend on its distinctive product. His commercial interest is clear: Turnstile sells precisely this kind of infrastructure. The reasoning nevertheless identifies a concrete cost that a quick demonstration can obscure: ownership continues after the first version works.
By September, his attention had moved to how finance teams budget for AI itself. Software costs tied to seats could be estimated through headcount. Software that performs work changes the calculation. He argued that spending now needs to be understood in relation to the work performed, with experiments earning larger commitments through measurable returns.
The conversation comes back to dinner
For someone working on reducing manual exchanges, Zamir still makes room for face-to-face conversation. He and Babineau described a Turnstile and First Round dinner with roughly 20 early-stage B2B founders, discussing how improvised founder effort becomes repeatable operating practice. Zamir said he had picked up ideas that could make him a better founder and CFO.
A June 2026 finance-leaders dinner in New York carried the same small-room approach. Zamir co-hosted with Seema Kumar and Edo Lawrence. The invitation promised conversation without a panel or slide presentation, with attendance capped at 20. The questions concerned billing models, reconciliation and whether companies could trust their revenue reporting. Even dinner had a relevant agenda.
Turnstile's next direction brings that trust question into work with AI agents. Its July explanation describes giving people and agents access to a shared commercial record through the Model Context Protocol, including read and write capabilities. Zamir's ambition is to make the underlying information coherent enough for new kinds of work to rely on it.
The thread through his career is responsibility for what an agreement sets in motion. He has examined the deal, helped operate the company behind it, and built software for the work that follows. A customer says yes. There is still a promise to keep, a bill to get right, and a number somebody should be able to explain.