Antoine Fort chose the sort of startup problem that can empty a room at a dinner party and fill a finance team's calendar: sales commissions. Every month, data arrives from several systems. Rules bend around territories, roles, quotas and exceptions. Someone tends a spreadsheet swollen with formulas. Someone else waits to learn what they earned. Then comes the message that turns a calculation into a small crisis of government: “My commission looks wrong.”
Fort, the co-founder and CEO of Qobra, saw a product hiding inside that ritual. His answer was software that pulls data from the systems where it already lives, lets operations and finance define calculation rules, shows salespeople where they stand and routes the result through approval before payroll. The mechanics are useful. The more interesting thing is what they reveal about his view of a company. Pay is never merely arithmetic. It is a promise, an instruction and a verdict, all compressed into one number.
A commission plan tells a salesperson which behavior matters. It tells finance how much growth will cost. It tells management whether its strategy can survive contact with a quarter. If the plan is obscure, late or wrong, distrust does not stay inside the spreadsheet. It walks around the office wearing shoes.
Three friends, three weeks, zero product
Qobra began in July 2020 with Fort and two fellow CentraleSupélec alumni, Tanguy Moullec and Axel Poitral. A friend told Fort about the Y Combinator habit of founding teams renting a house and building together with theatrical intensity. The trio borrowed the useful part of the myth. They went to the Mediterranean for three weeks and coded and designed wireframes without interruption.
Fort later described their opening balance with cheerful severity: zero customers, zero product, zero investors. Seven business angels then put in €100,000. Indy, Partoo, Welcome to the Jungle and Didomi became early customers. Revenue brought the founders their first salaries and gave the software something more valuable than elegance: consequences. Real compensation plans contain oddities that a pristine demo never will. The product had to survive them.
The founders' mix made sense. Fort had studied engineering and computer science after preparatory work in mathematics and physics. Before Qobra, he moved through data science at Cuestionarix, business development at The Triana Group in New York and digital transformation consulting at Sopra Steria. The sequence reads like an accidental syllabus for enterprise software: understand data, learn to sell, watch organizations resist change, then build.
The spreadsheet was only the visible enemy
It is tempting to describe Qobra as a campaign against Excel. Fort's sharper complaint is about invisibility. A spreadsheet can calculate beautifully and communicate terribly. It may update once a month, live on one person's laptop and offer a salesperson no clean line from a signed deal to a future payout. In that fog, even a correct answer can feel suspicious.
Qobra's workflow begins upstream. It connects with customer-relationship systems, planning tools and data warehouses. Operations teams define rules in a no-code interface. Reps see performance and projected commission in near real time. A validation sequence gives managers, finance and human resources a chance to review before the money moves. This turns the commission statement from an archaeological report into a live instrument panel.
Fort argues that variable pay is usually the largest go-to-market investment in a business-to-business company. That makes the commission plan a management document disguised as a formula. A company that rewards new revenue, retention or margin will receive different behavior. A plan with too many clauses may technically encode the strategy while ensuring nobody can remember it before lunch.
His public advice tends toward legibility: set objectives that fit each role in the sales cycle, communicate the plan, keep performance visible and avoid capping the people who outperform. He also emphasizes belonging. Sales performance, he has said, is more than a pile of individual results; mutual support and collective success matter too. This is a less mechanical idea than one might expect from the founder of calculation software. The formula sets a direction, but the team still has to travel together.
Capital for the unglamorous work
In 2022, Qobra raised a €5 million seed round led by Breega. Fort described the purpose plainly: at five people, the team needed to become a company, hiring across technology and sales to meet demand it could no longer absorb. The next round had a different job. In October 2023, Qobra announced a €10 million Series A led by Singular, with Revenue Syndicate and existing investor Breega participating.
The money was intended for deeper product capability, larger customers and international growth, particularly in Britain and the United States. Qobra opened its London office in June 2024. By then, the platform served companies in multiple countries and across industries far removed from the startup cliché: software, property, financial services, insurance, pharmaceuticals, medical devices and automotive businesses all have salespeople whose incentives eventually become somebody's spreadsheet.
Fort, Moullec and Poitral found Qobra after a three-week Mediterranean build sprint.
A €5 million seed round funds hiring across the young company.
A €10 million Series A backs product development and international expansion.
Qobra opens in London; Revenue Connect gathers more than 250 operations professionals.
Fort pushes Qobra and his own workflow toward supervised AI agents.
Fort's stated entrepreneurial principle is not exotic. Do good work, repeat it each day and do not give up. Founders are often encouraged to produce a philosophy grand enough for a conference stage. His is closer to an instruction printed above a workshop bench. The interest lies in its repetition. A correct payout earns trust once. Hundreds of correct payouts turn reliability into reputation.
The founder builds an agent
Growth created a fresh version of the visibility problem. As Qobra moved beyond 50 employees, Fort wrote that he was losing sight of small but important signals: customer satisfaction, feature requests, sales objections and the quiet measures scattered along the customer journey. The company had capable operators and regular meetings. He still missed the details a founder remembers because he helped construct each corner.
His response was characteristic: connect the systems. In roughly four hours, he built an AI chief of staff drawing from Salesforce, Modjo, Linear, Slack, Gmail and Intercom. Each morning it delivered three priorities, notable deals with recent customer meetings and clients that might need attention. The tool mattered, but so did the order of operations. Fort wanted the company to become AI-native and believed he should change his own routine before asking the team to change theirs.
That experiment now mirrors Qobra's product direction. The company has presented three supervised agents: an Architect to turn compensation documents into plans, an Analyst to interrogate commission data and a Coach to answer reps' questions. The ambition moves beyond automating formulas. It makes the system conversational while preserving human review. In compensation, an answer delivered quickly is useful; an answer that can also show its work is safer.
There is an appealing symmetry here. Fort began by attacking a monthly information delay between a company and its salespeople. Six years later, he is attacking the information delay between a growing company and its founder. Both projects ask the same question: which facts must arrive, for whom, and soon enough to change a decision?
A company is what it makes clear
Fort has said Qobra wants to become a global commission-management platform. Funding, offices and product modules provide the conventional markers of that attempt. The deeper continuity is trust. A salesperson should understand how effort becomes pay. Finance should understand how performance becomes cost. Management should be able to test whether an incentive produces the behavior it intended.
None of this removes disagreement. A transparent plan can still be stingy, misguided or badly timed. Clarity merely makes the argument honest. That is valuable. Organizations accumulate power in their opaque systems: the private spreadsheet, the unexplained exception, the metric that arrives too late to contest. Turning those systems into shared interfaces does not make work fair by magic. It makes fairness easier to inspect.
This is why Fort's unglamorous problem travels. Commission management sounds narrow until it is seen as a test of organizational character. Does the company say what it values? Does it calculate what it promised? Can an employee see the answer before the answer becomes irreversible? Software cannot supply virtue. It can, however, make ambiguity much less comfortable.
The original Qobra trio started beside the Mediterranean with wireframes, code and no customers. The scenery was cinematic; the work was bookkeeping. That contrast still suits the company. Behind every triumphant gong in a sales office waits a decimal, a rule and an approval. Fort's wager is that handling them well can turn administration into motivation. The future of work may occasionally arrive wearing sunglasses. More often, it arrives as a number everyone finally agrees is right.