The Fuzu Story: How two Finns turned one Nairobi driver hire into fifteen hundred CVs, outlasted a decade of African online recruitment failures, and quietly walked into the machinery of the AI industry
In April of twenty twenty six, in an industrial neighborhood on the edge of Nairobi, more than eleven hundred people walked out of an office building for the last time. They had spent their days looking at things nobody wants to look at. Violent images. Hate speech. Text that a machine could not be trusted to judge alone. They had labelled that content so that other software, running on servers in California, could learn what was safe and what was not.
Their employer was a company called Sama. It had run some of the most notorious contracts in the short history of the generative AI industry, including the labelling work that made the first version of ChatGPT publishable. The workers who had done that job had already sued. Then, in April, Meta ended a major moderation contract, and eleven hundred and eight jobs in Nairobi disappeared in a single announcement.
A few kilometres away, in a much smaller building, a Finnish Kenyan company called Fuzu was quietly doing the same category of work, and quietly building a very different version of it. For eleven years Fuzu had been an online jobs marketplace. Now, months after that rival’s mass layoff, Fuzu was surfacing a new division called Atlas, selling managed African teams to Microsoft, to Mastercard, to global clients across the artificial intelligence industry. It was a pivot the founder himself refused to call a pivot.
This is the story of Fuzu. The company built from a single moment when fifteen hundred people applied for one Nairobi driver job. The company that outlasted a decade of African online recruitment failures on less capital than most of its rivals had already lost. And the company that read the disruption of its own category by generative AI, and decided to move into the machinery of generative AI itself.
THE ORIGIN
The origin of Fuzu does not begin with a jobs marketplace. It begins with a driver. In twenty ten, Nokia was still the largest handset maker in Africa. Its research centre in Nairobi, opened two years earlier, was the largest single research and development investment by any Western technology company on the continent. That year, Nokia was hosting an open innovation conference in Nairobi, and a Finnish vice president named Jussi Hinkkanen was responsible for the logistics. The conference needed a driver for VIP transportation. The team posted the job. Approximately fifteen hundred CVs came back.
Fifteen hundred CVs, for one low skill role. There was no way to read them. No way to verify them. No way to rank them. No way even to call the applicants back. Somewhere in that pile was the right person, but the infrastructure to find that person did not exist. Hinkkanen and his colleague Jussi Impiö, the founder of Nokia’s Nairobi research centre, spent the next two to three years turning that gap into a thesis. The African labour market was not short of talent. It was short of the machinery to match talent to opportunity at scale.
Hinkkanen was not a stereotypical technology executive. He had studied industrial economics in Finland, then spent much of his early career in Mozambique, advising the ministry of science and technology, running the country’s research and education network, working with the United Nations Development Programme, and leading a thirty million dollar World Bank project on electronic government infrastructure. He had built things in African economies before he ever sold a phone. When Microsoft acquired Nokia’s handset division, he stayed through the transition. Then, in twenty thirteen, he quit Microsoft entirely. He and Impiö incorporated Fuzu in Helsinki. The company would be Finnish on paper, Kenyan in practice, and dual hub from day one.
For two years the company built quietly. Hinkkanen took the CEO title in twenty fourteen. The team hired a Kenyan operator named Robert Kimani, who had until then been the chief executive of BrighterMonday, one of the leading incumbent Kenyan job boards. That single hire was a statement. Fuzu’s first senior Kenyan leader came out of the exact category Fuzu was about to disrupt. On the seventeenth of September, twenty fifteen, Fuzu launched publicly at the Fairmont Norfolk Hotel in Nairobi. Guests included the Rockefeller Foundation, the mobile phone financier M-KOPA, and Equity Bank. Within forty days the platform had thirty thousand Kenyan users.
The category Fuzu had chosen was already, visibly, a graveyard. The Swiss publisher Ringier owned BrighterMonday in Kenya and Jobberman in Nigeria under a holding company called One Africa Media. In December of twenty sixteen, only fifteen months after Fuzu launched, Ringier shut BrighterMonday Rwanda and Cheki Tanzania, and closed operations in Zimbabwe and Zambia. Kenyan startup Kuhustle, launched the year before Fuzu, had already pivoted away from its original model after quality of service complaints, and would never scale. Across the continent, the online jobs category was old, unloved, and losing money. Into that graveyard, Fuzu was about to double down.
THE STRUGGLE
To understand what Fuzu was up against, understand the economics of an African jobs marketplace. Job seekers get everything for free, because charging them means locking out the exact audience the platform is supposed to serve. So the money has to come from employers. And African small businesses have infinite substitutes. They can post on LinkedIn. They can drop a message into a WhatsApp group. They can ask their church. They can put up a notice at the corner shop. Meanwhile every online listing competes with a huge secondary economy of WhatsApp job scams that has trained job seekers to distrust every offer they see. That is the terrain. Charging small employers for something they can already do informally, in a market where jobs are as likely to be fake as real.
Fuzu’s answer was to build a product wide enough to survive that pressure. Job seekers got a free profile, a free skills assessment, a free personality assessment, a CV builder, a learning marketplace called Fuzu Learn, and a community forum where the founder himself ran public question and answer sessions. Employers were charged in three tiers. A starter plan from around sixty dollars per job posting. A pro plan from three hundred dollars per month for teams that hired regularly. And a top tier called elite, from six hundred dollars per successful hire, with a dedicated recruiter and a replacement guarantee. Everything on the platform was designed to make the free product so good that employers would eventually pay for the pipe of talent flowing through it.
In December of twenty sixteen, Fuzu closed its first institutional round. One point seven five million euros. The lead was Finnfund, the Finnish development finance institution. Alongside came the Finnish staffing group Barona, and a Finnish angel vehicle called Polkuni. By the close of that round, roughly one hundred and eighty thousand Kenyans had registered on the platform. By late twenty seventeen, that number had crossed half a million. In twenty eighteen, Fuzu launched in Uganda. The team was around twenty six people, most of them in Nairobi, a handful of engineers in Helsinki.
While Fuzu grew, its peers kept dying. Kuhustle never recovered. Lynk, the Kenyan informal services marketplace, was eventually sold in twenty twenty two to a Nigerian startup called Eden Life in what was widely read as a distressed exit. Ringier consolidated its remaining African job boards under a new brand called the African Talent Company, propped up by a Mastercard Foundation grant subsidy. The South African cleaning services marketplace SweepSouth raised eleven million dollars from investors, entered Kenya and Nigeria, and withdrew from both within months, citing global macroeconomic conditions. Every one of these was a company that had raised more money than Fuzu, launched louder than Fuzu, and lost. Fuzu’s edge was that it had never tried to grow at the pace those companies had tried to grow. It was small. It was patient. And it was still there.
THE PIVOT
The first real turn came on the tenth of March, twenty twenty. Fuzu closed a Series A of three point four million euros, roughly three point nine million dollars. The lead was Sparkmind, a Nordic education technology fund. Finnfund followed on. So did Cornerstone Enterprises in Kenya, the Japanese corporate vehicle Aucfan Incubate, the Swiss firm Seedstars International, and Kepple Africa Ventures out of Japan. The plan was clear. Build artificial intelligence into the recruitment product. Launch Nigeria.
The Series A closed just as the world began to lock down. Kenya, Uganda, and Nigeria all went into some form of restriction. Every startup on the continent had to make a decision about how much of its runway to burn preserving jobs versus preserving cash. Fuzu is not known to have run public layoffs during this period. Its name does not appear in any of the layoff trackers that captured the sector’s pain. Instead, the company kept building. In twenty twenty one, it launched in Nigeria, hiring a country director named Patricia Duru to lead the market. In twenty twenty two, it released a mobile app. Internally, an operator named Muthoni Mathenge led an overhaul of the business model that Finnfund would later credit with increasing revenue by fifty percent.
In September of twenty twenty two, Fuzu topped up its balance sheet again. This time Finnfund put in five hundred thousand euros as part of a broader mezzanine facility, alongside Sparkmind. This was not a priced equity round. It was a bridge. And it would turn out to be the last piece of new institutional capital Fuzu would take for years. Then, in twenty twenty three, something quietly appeared on the company’s product roadmap. It was called Fuzu Global Workforce. On the surface it was a way for European and American employers to hire African talent remotely. Underneath, it was the seed of something much larger. Fuzu had begun building the machinery of a services business, not just a marketplace.
While Fuzu built quietly, the rest of the African jobtech industry was collapsing loudly. In March of twenty twenty four, Remotasks, the annotation arm of the American company Scale AI, abandoned Kenya, Nigeria, and Pakistan almost overnight, stranding thousands of gig workers who had built their income around the platform. Around the same time, Sama was still absorbing the fallout of a lawsuit brought by a former content moderator named Daniel Motaung, followed by a class action by more than a hundred and forty former Meta moderators. All of these were companies that had chosen to compete on volume and price. Fuzu, watching, was choosing the opposite path.
THE SCALE
In early twenty twenty six, Fuzu Global Workforce was rebranded as Fuzu Atlas. The new positioning was direct. Atlas was a governed AI data operations division. Not a gig platform. Not a microwork marketplace. A managed service that ran multilingual data operations, evaluation of large language models, safety review, multimodal annotation of images and video, and audit oriented expert oversight for regulated use cases. Delivery pods were live in Nairobi, in Lagos, in Cairo, in Sofia, and in Buenos Aires, with Helsinki as the group headquarters. The vocabulary was chosen with care. Every marketing page used the words governed, expert in the loop, and compliance aware. Those words were, quite deliberately, the opposite of everything Sama had been criticised for.
The client roster Fuzu was willing to name publicly was blue chip. Microsoft. DHL. Heineken. Airtel. Mastercard. Unilever. The company said it had worked with more than two thousand organisations. What is worth noticing is what was not on that list. No frontier AI lab. No OpenAI. No Anthropic. No Meta. No Scale AI. Either the lab side work is under confidentiality, which is standard in this industry, or Fuzu’s real book of business is enterprise clients doing artificial intelligence adjacent work, rather than the training of frontier models. The truth is probably somewhere in between, and the company has never clarified it.
Then, in the week of the sixth of July, twenty twenty six, two things happened at once. The Jobtech Alliance, an initiative of Mercy Corps and BFA Global, announced a catalytic investment in Fuzu. The size was not disclosed. A companion investment in a Kenyan retail startup called Kyosk was announced the same day. The Jobtech Alliance had, one year earlier, published a sector wide research paper arguing that generative AI was gutting the low end of African digital work, and that the resilient future belonged to what it called demand led managed service platforms that embed quality control, accountability, and human oversight. That paragraph read like a description of what Fuzu Atlas had just publicly launched.
At the same time, the Fuzu Atlas landing page went public. And this is where the pivot got interesting. The founder, Hinkkanen, refused to call it a pivot. His public posts on the company’s tenth anniversary in twenty twenty five spoke only of continuity, of the people who had built the company, of the next chapter of the future of work. The word pivot came from outside. It came from Nick Markham, the Jobtech Alliance investment lead, who told the press, quote, most African talent platforms still operate as marketplaces, competing on volume and cheaper sourcing. Fuzu has already moved past that, end quote. The founder spoke continuity. Third parties spoke transition. That gap was not an accident.
TODAY AND TOMORROW
Which brings us to today. Fuzu, in twenty twenty six, is two companies living inside one. On one side sits the original marketplace, still running in Kenya, in Uganda, in Nigeria. Fuzu’s own about page continues to call the platform Africa’s number one career platform, and job seekers on the site would notice no difference. On the other side sits Atlas, the higher value services business, quietly signing contracts with global enterprises and slowly building out an alternative to the low wage model that has defined African AI data work for a decade.
The financial picture is unusually thin for a company this old. Roughly five and a half million euros in disclosed institutional capital across eleven years. No priced round since March of twenty twenty. No valuation ever published. No revenue ever disclosed. That absence of a fresh priced round in the years around the pivot is itself a signal. Either Fuzu chose not to raise, or the market did not offer terms Fuzu was willing to accept. Either way, the Atlas pivot is not a bet made from a position of abundance. It is a bet made because the classic marketplace path, as every peer around Fuzu has just proven, no longer leads anywhere good.
The open questions are the ones that will define the next five years. Can Atlas win real frontier AI work, or will it remain an enterprise vendor by another name. Can Fuzu preserve the marketplace’s public trust while treating it, internally, as a talent funnel for a very different business. And can a company that spent a decade selling access to jobs in Kenya, in Uganda, in Nigeria, credibly sell the humans who train the models that will decide which of those jobs still exist in ten years time. The bet is unresolved. The pivot is underway. And unlike almost every peer that surrounded Fuzu when it launched, Fuzu is still there to make it.
In twenty ten, a Finnish executive tried to hire one Nairobi driver and received fifteen hundred CVs. The answer to that problem became a company that survived a decade of African jobs marketplace failures on less capital than most of its rivals had lost, and then, in twenty twenty six, quietly reinvented itself into the machinery of the artificial intelligence industry that was gutting its original category.
This is Asili Africa. Every empire has an origin.
Key Takeaways
- THE ORIGIN. The origin of Fuzu does not begin with a jobs marketplace. It begins with a driver.
- THE STRUGGLE. To understand what Fuzu was up against, understand the economics of an African jobs marketplace.
- THE PIVOT. The first real turn came on the tenth of March, twenty twenty.
- THE SCALE. In early twenty twenty six, Fuzu Global Workforce was rebranded as Fuzu Atlas.
- TODAY AND TOMORROW. Which brings us to today. Fuzu, in twenty twenty six, is two companies living inside one.
Listen to the Episode
Also available on YouTube — search “Asili Africa” or subscribe to our channel.