He spent a decade keeping SaaS integrations alive by hand - 50 connectors, 1,000 customers, 15 people. Instead of complaining, he built the thing that makes that grind obsolete. It's called StackOne.
Most founders chase a market. Romain Sestier chased a scar. For the better part of a decade, at one software company after another, the same wound kept reopening: integrations that broke overnight, drained engineering teams, and quietly slowed everything down. He kept them running anyway. Then, in 2023, he decided the smarter move was to make the whole painful ritual unnecessary.
The result is StackOne, a platform that lets B2B SaaS products and, increasingly, AI agents connect to hundreds of enterprise systems through a single unified API. If that sounds like plumbing, that is exactly the point. Sestier has spent his career in the parts of software that nobody puts on a highlight reel, and he built a company on the conviction that the unglamorous layer is where the leverage hides.
Sestier did not come to integrations through computer science. He studied Business Management and Marketing at Royal Holloway, University of London, and then walked straight into the operational trenches of B2B software. At Maxymiser, later acquired by Oracle, he crossed paths with a solutions engineer named Guillaume Lebedel, whose job was building integrations. The two would keep working near each other for the next ten years.
The pattern repeated everywhere he went. At ContentSquare he opened the UK office and ran services. At Yieldify, later acquired by Publicis, he became VP of Product and Customer Success, managing a global team of around 70 and overseeing an integration strategy that, at its most stretched, meant keeping roughly 50 connectors alive for 1,000 customers with a team of 15. Every job had a slightly different logo on the door and the same headache waiting inside.
That last stop matters. From 2020, Sestier spent time at Google, first as a co-founder and product lead inside Area 120, the company's in-house incubator, and then as a lead product manager for Ads Sales Intelligence. There he helped scale an AI-driven insights product that delivered around $1 billion in incremental revenue. It was, by any measure, a career-defining win. He left it anyway.
The obvious question is why the integration problem, old and familiar, suddenly became worth a company in 2023. Sestier's answer is timing. For years the integration layer was something teams bolted onto a product after the fact, a tax you paid to sell to enterprises. Large language models, he argues, made it possible to reimagine that layer from scratch.
The insight is subtle. StackOne does not just let a human developer ship a connector faster. It uses models to help write deterministic, enterprise-grade integrations, and it exposes those integrations in a way that an autonomous AI agent can actually use to take action across the systems a company runs on. As Sestier frames the new positioning, StackOne is the integration infrastructure AI agents need. Agents can reason and plan; the missing piece is the ability to act across Workday, Salesforce, and 450 other places where real work lives.
When Sestier finally started StackOne, he did not go looking for a co-founder. He already had one. Guillaume Lebedel, the solutions engineer from the Maxymiser days, became CTO. A decade of working near the same problem, across the same companies, meant the two did not need to align on what they were building. They had both bled on it.
StackOne reached six-figure revenue within weeks of launch. That speed is less a growth-hack story than a symptom: when a founder has personally felt a problem dozens of times, the solution tends to sell itself.
Their shared framing for the whole venture is disarmingly plain, and Sestier repeats it often.
StackOne began with a $3.6M seed from Episode 1 Ventures, a nod to the company's London roots - its main contact line still carries a UK dialing code. In May 2025 it raised a $20M Series A led by GV, bringing total funding to roughly $23.6 million. The more telling detail is the shape of the cap table.
Around that table sit GV, Sequoia, GitHub, XTX Ventures, Workday Ventures, and, notably, both OpenAI and Microsoft. It is not every day that two of the fiercest rivals in AI show up backing the same infrastructure bet. Workday's presence is its own signal: one of the enterprise systems StackOne connects to is also, through its venture arm, an investor.
Ask what Sestier carried out of Google, and the answer is not the $1B headline. It is Area 120's culture of moving fast, experimenting constantly, and making bold product bets. StackOne runs on bi-weekly sprints and a willingness to ship. Yieldify taught him to tie a roadmap to revenue outcomes; Google reinforced how much user experience and adoption metrics matter. The blend shows up as a product team that ships quickly but measures ruthlessly.
The other inheritance is a hard line on trust. StackOne is built around real-time data transfer with zero data storage, and it carries SOC2 Type II, HIPAA, GDPR, and CCPA compliance. For a company asking enterprises to route sensitive HR, recruiting, and CRM data through a single pipe, that posture is not a feature list. It is the price of entry.
A marketing graduate now leads a deeply technical developer-tools company whose stack spans Rust, TypeScript, Python and Ruby on Rails - and which publishes open-source MCP servers so AI agents can plug in.
Sestier sizes the opportunity at a $280 billion market, resting on a simple observation: more than 90% of B2B SaaS products need to offer integrations, and almost none of them want to build that muscle in-house. His ambition is to make StackOne the default way any AI agent or SaaS product securely acts across every enterprise system in real time.
Whether he gets there is unknowable. What is clear is the through-line. From maintaining 50 connectors with a skeleton crew, to shipping a billion-dollar product inside Google, to a cap table stacked with rivals, the same instinct keeps showing up: find the layer everyone else treats as a chore, and rebuild it into the thing everyone depends on. Sestier spent ten years living the problem. Now he is betting the rest of the decade on the fix.