In 2019, Mehdi Boudoukhane and Thibaud Elzière were sketching a product in Brussels with a rather inconvenient target: Atlassian. They wanted a different way for product teams to work. Six years later, Atlassian acquired Cycle’s technology and team. The rival had become the destination. Boudoukhane’s account of the deal includes another detail: it closed on the day his daughter Emna was born. Some days are difficult to fit into a calendar square.
The acquisition was announced on September 3, 2025. It makes an appealing ending, with the symmetry usually supplied by a screenwriter. Yet Boudoukhane’s story is more interesting when the ending is allowed to remain untidy. He began with a broad ambition, spent years learning how to make it useful in a narrow setting, and joined a larger organization to keep working on the same problem.
That problem sounds almost comically modest. When a customer asks for something, can the people building the product hear it? Can they connect that request to a decision? And when the work is finished, can anybody remember to tell the customer? There is enough room between those questions for an entire software company to disappear.
The occupational hazard of asking why
Boudoukhane started as an engineer and data scientist. His habit of asking why eventually prompted his CEO to suggest product management. The role put him among designers, engineers and customers, with responsibility for understanding what should be built as well as how. His education had taken him through the Université libre de Bruxelles between 2008 and 2013; his subsequent work included Riaktr and Lab Box.
He liked the intersection of technology, design and business. He was less taken with the number of tabs required to stand at it. Designers had Figma. Engineers had GitHub. Product managers had a collection of tools that still left them assembling the story themselves. More software had produced more places to look.
Before Cycle, he tried stitching together a solution with Airtable, Notion and Zapier. That is a familiar temptation for someone who can build: perhaps one more connection will make the whole arrangement behave. His frustration survived the connections. In 2019 he teamed up with eFounders, the startup studio now called Hexa, to build a dedicated product. Its working name was ProductX, a name with the useful property of promising almost anything.
The early idea was expansive. Product teams would share a flexible workspace, with collaborative documents at its center. A document could hold the context that an issue or ticket tended to squeeze out. The platform would fit different teams’ terminology and connect to their existing tools. Boudoukhane wanted a product that people could make their own without having to assemble it themselves.
Fifty conversations before the answer
Boudoukhane flew to San Francisco and interviewed more than 50 product leaders, including people at Dropbox, Figma and Superhuman. The travel had a specific purpose. If he wanted to build for demanding product teams, he needed to watch them work. An ambition can look perfectly sensible on a Brussels whiteboard and rather different across a table from someone who must use the result.
He came back with a clearer picture of the customers he wanted. On a second trip, he announced Cycle’s beta in a street video. There were investor meetings and customer conversations, followed by evenings working through the implications around an Airbnb kitchen table. The company was still young enough that a change of accommodation could double as a change of office.
The early prototype took a month to ship. Its first user, Cyril from Folk, put up with the bugs. Boudoukhane has kept the first wireframes, an unusually honest souvenir for a founder who cares about design. They show the distance between a convincing explanation and a product another person can comfortably use.
One encounter stayed with him. At Superhuman, he saw a team telling users when features they had requested were released. The work was careful and laborious. It also made the relationship visible: someone had asked, someone had listened, and someone had come back with an answer. In his later telling, this became a picture of the future he wanted to build toward.
A smaller promise that people could use
Cycle stayed in beta for almost three years. Boudoukhane’s retrospective explanation is blunt: the team was trying to build the full future before it had won a first use case. Each addition seemed consistent with the vision. Taken together, they made it harder to arrive at something customers could immediately understand.
The eventual focus was customer feedback: gather it, connect it to the product work, and close the loop after a release. That gave Cycle a clear job inside an organization. A support conversation or a sales call could become evidence for a product decision; the original customer could remain attached to that evidence.
In July 2022, Cycle’s Product Hunt launch became Product of the Week. By its February 2023 funding announcement, the company reported that active companies had grown from 50 to 500 after that launch. It also reported that more than 70% of users worked outside product management. Designers, engineers, founders and customer-facing colleagues were participating in a process often described as the product manager’s job.
Company-reported figures, February 2023. A snapshot of the early product, not current usage.
The funding announcement put $6 million behind the experiment. Boldstart led the seed round, joined by funds including Base Case, The 20VC Fund and SV Angel, alongside 60 operator angels. Among them were product practitioners such as Shreyas Doshi and Scott Belsky. Boudoukhane had gone looking for people who understood the work; some became backers as well as interlocutors.
He also imagined Cycle as a future record of a product person’s contributions, much as GitHub makes an engineer’s work visible. That aspiration went beyond organizing requests. It concerned who gets recognized for the work between an idea and a release, a stretch of territory where credit can be as scattered as feedback.
The door that needed a person beside it
The product’s flexibility brought a commercial complication. An empty workspace could take time to configure. Feedback lived in external systems, and a new user might lack permission to connect them. There was no guarantee that five minutes alone with the software would reveal its value.
Boudoukhane changed Cycle’s activation process. Prospective customers booked an onboarding call, received help setting up a trial and discussed pricing in context. For a founder who loved products, adding a human conversation to the entrance was a consequential choice. It meant accepting that an elegant interface could still need an introduction.
His February 2024 essay on product-led growth attacks the habit of turning a sales strategy into a company identity. He separates retention, activation and acquisition, arguing that they can need different approaches. The practical question is whether a customer gets value. A fashionable acronym is poor compensation for a workspace nobody knows how to set up.
The admission gives his public writing some bite. He includes himself among the founders who stayed too long in familiar territory, trying to solve commercial problems by building more product. The account has a recognizable engineer’s embarrassment: the mechanism was working, and the person still could not get through the door.
- 01Hear itCapture feedback where customers already speak.
- 02Connect itLink customer context to product decisions.
- 03Return to itTell people when their request becomes a release.
An AI test with an impolite name
Cycle spent its first three and a half years without an AI feature. When the team adopted AI, it had a task already waiting: people were spending hours processing feedback manually. The opportunity lay inside a working process, among the sorting, extraction and connections needed to make customer conversations useful.
In June 2024, Boudoukhane published what he called the “Bullshit AI” test. His questions ask whether a product has value without AI, whether AI improves its central job substantially, and whether the same result is easily achieved with prompts in a separate chat window. The tenfold improvement he proposes is a test of ambition, rather than a measured performance claim about Cycle.
He applies the criticism to his own company. Cycle’s first AI feature was basic content generation, precisely the kind of addition he later found unconvincing. The more meaningful work was automating feedback processing. It is pleasing to encounter a detector whose inventor remembers to point it at himself.
His later writing returns to a related mistake. Cycle’s prototype accumulated features faster than structure, until the team rebuilt it from scratch. In 2026, he argues that cheaper code generation makes prototypes abundant while coherent products still require judgment. Someone must decide how the pieces fit, which changes belong and what the system should be able to withstand.
“The hard part of software was never code.”
Mehdi Boudoukhane, The Quiet Force, March 2026
The rival becomes the next workplace
Atlassian’s acquisition brought Cycle’s team into the organization behind Jira Product Discovery. The September 2025 announcement described a familiar problem: product managers piecing together interviews, support tickets, sales feedback and community discussions. Cycle’s technology would help turn those fragments into customer signals connected to roadmap ideas.
The match was specific. Boudoukhane had spent years working on the information that reaches a product team, and the information that goes back out. Atlassian wanted those capabilities alongside its discovery workflow. The original challenge could continue within the organization it had named as a benchmark.
By 2026, his public profile described him as Principal Product Manager for Feedback at Atlassian. He also writes The Quiet Force and advises founders, including the team behind Enobase. The titles have changed; his writing still worries at the same distance between a persuasive idea and something people can use.
In an April 2026 essay, he describes startups as a sequence: a first problem you can solve, an expansion that follows, and the larger future you intend to reach. His own story provides a slightly mischievous illustration. He wanted to build the next Atlassian. To get moving, he had to make a smaller promise about customer feedback. Then Atlassian bought the team that had learned to keep it.
The useful ending belongs to the customer. A request enters a support conversation, finds its way into a decision, and returns as news of something shipped. Boudoukhane’s six-year detour was devoted to making that journey less accidental. Inside Atlassian, he is still working on the return trip.