Arkestro orchestrates the buy-side.

The buyer who stopped opening Excel

On a Tuesday morning at a Fortune 500 manufacturer somewhere off the I-680, a sourcing manager opens her inbox and finds 14 supplier responses already scored, ranked, and flagged for the one bid that looks suspiciously low. She did not build a pivot table. She did not chase a quote. She makes one decision before lunch. That decision used to take six weeks. It now takes ninety minutes. Arkestro is the reason - though most of the people inside the company never see its logo. The platform runs quietly, doing the dull and decisive math that procurement teams have always done by hand, only faster and with fewer regrets.

Arkestro calls this category "predictive procurement," which sounds like a phrase invented at a conference. It was. The category, however, is real. So is the savings claim - 18.8% on every million dollars of spend, according to the company's own customer data, audited by procurement analysts who are professionally skeptical.

The problem nobody wanted to solve

Procurement is the department companies forget exists until something stops arriving. It is also, by spend, the largest cost center in most enterprises. The math is unkind: a 1% reduction in purchased-goods cost typically beats a 10% increase in sales. And yet, until embarrassingly recently, the people running enterprise procurement were doing it in spreadsheets, hand-cleaning supplier responses, and emailing PDFs back and forth across three time zones.

The reason is structural. Buying things at scale involves thousands of line items, hundreds of suppliers, and a thicket of contractual terms. Software exists - SAP Ariba was the original, and it has been the category leader for twenty-five years - but the software was built to record the work, not to do it. So the work stayed manual.

This is the irony Arkestro built itself around: the very technology meant to liberate buyers turned them into the world's most expensive data-entry clerks. The category was overdue for a rude awakening. It arrived in 2016, calling itself BidOps.

The founders' bet

Edmund Zagorin was a procurement consultant before he was a founder, which is a polite way of saying he had watched the disease up close. He kept seeing the same pattern - smart sourcing professionals using their week to do the work of a script. He persuaded a childhood friend, Ben Leiken, then a product leader at SurveyMonkey, that this was a software problem hiding inside a workflow problem. Leiken became the CTO.

The bet was twofold. First: most of what buyers want is predictable from data the company already has. Therefore, pre-fill it. Second: sourcing events are repeated games between buyers and suppliers, which means game theory applies. Therefore, model it. Wrap both ideas in an interface a buyer will actually open and you have a product.

The bet got a public-relations boost when Rob DeSantis, an original co-founder of Ariba, the very company Arkestro was implicitly arguing had run its course, decided the upstart was right. He interviewed the customers, the engineers, and the investors, and then joined as a co-founder. It is the kind of endorsement money cannot buy, because money rarely tries.

What the software actually does

For a product that has raised over $100 million, Arkestro's user-facing surface is suspiciously calm. The Predictive Procurement Platform does three things, in roughly this order: it prefills, it predicts, and it scores.

The pitch is unglamorous, which is the point. Buyers do not want a chatbot. They want fewer surprises and a defensible award. Arkestro built around that quietly correct premise, and the customers - most of whom decline to be quoted on the record because their CFOs would rather not advertise the savings - kept paying for more seats.

The proof, with bars

Arkestro's published numbers cluster around a single, defensible claim: buyers who use the platform save more than buyers who do not. The savings vary by category, and the company is unusually candid about the variance. Oil and gas tends to outperform. Indirect services tend to lag, because the data is messier.

The customer roster is, by design, hard to read. Fortune 500 manufacturers, energy firms and global logistics operators do not put their procurement vendors on billboards. The investor list is louder: NEA, Koch Disruptive Technologies, Altira Group, Aramco Ventures and Activant Capital are not in the habit of mistaking a category fad for a real one.

The mission, minus the slogan

Arkestro's leadership is allergic to procurement-industry maximalism. They do not promise to "reinvent" the function. They promise to make it faster - which, when the function is responsible for trillions of dollars of global spend, is a more ambitious claim than reinvention. The mission, stripped of its marketing varnish, is to turn enterprise procurement from a craft into an engineered system.

That has cultural consequences. The company is remote-first, technical, and heavy on data scientists who can read an RFP without their eyes glazing over. The hiring bar leans toward people who have lived inside a procurement department - not consultants who flew in for a quarter. It is the kind of organization that takes its category seriously enough to write internal style guides about the difference between "sourcing event" and "procurement event," and then enforces them.

Why it matters tomorrow

Supply chains have spent the last six years being publicly humiliated. Pandemic shocks, shipping crises, geopolitical reroutes and the slow grind of inflation have taught every CFO in the country to care about what their procurement team does. That attention does not go away. The next wave of pressure - tariffs, energy transitions, AI-driven demand spikes - will fall on the same buyers, with the same spreadsheets, unless something changes.

This is the tailwind Arkestro is riding, and the reason Aramco Ventures wrote a check. The world's largest energy buyer does not invest in workflow software for fun. It invests because the workflow is the bottleneck, and bottlenecks compound.

There are competitors - Coupa, Keelvar, Fairmarkit, the entrenched Ariba - and the category will not be a winner-take-all. But Arkestro has a structural advantage few of them do: it was built, from the first commit, to make a prediction before the buyer asks. Everything else is an interface around that bet.

Back at the Fortune 500 manufacturer off I-680, the sourcing manager closes her laptop at 4:42 p.m. The award is signed. Three suppliers were rerouted. The bid that looked too low turned out to be a clerical error - the platform flagged it, she fixed it, no one lost a contract over a typo. She did not build a pivot table. She did not chase a quote. She made twelve decisions before lunch, and the rest of her afternoon belonged to her. That, more than any savings percentage, is the thing Arkestro built. The pivot table is dead. The buyer went home.

Footnotes and curiosities