When I was 9, my dad lost his job and slipped into long-term unemployment. It was 1994. The Franc CFA, the currency we used in Cameroon, had been devalued overnight, reducing my dad’s productivity value by 50 percent due to decisions made 6,000 kilometers away in Paris. I couldn’t fully grasp what was happening, but I knew it was bad.
Extreme lack of agency leads to a multitude of coping mechanisms. My dad chose alcohol, with all its consequences. Fast forward to 2001, I was a fresh high-school graduate, ready to conquer the world. I moved from my small hometown, Bafoussam, to Yaoundé, the capital city, for university studies. And I realized: small town, small problems, big town, big problems. But also more opportunities.
Due to the overcrowding and beautiful chaos of my streetless neighborhood, it was impossible for even a bike to find its way through walkways running from someone’s kitchen to another’s toilets. So, every single early morning, I would pick up bags of hot bread along the closest passable road to ensure last-mile delivery for the mom-and-pop stores inside our hood. The hustle was good, but I was still wondering what to do with my life.
Eventually, I had to find my luck somewhere else. I landed in Germany, a country whose language I did not speak, with no friends, no family, and no easy way back. At the time it felt like I was playing Russian roulette with my life. But I ended up getting my PhD. It turned out to be the better decision compared to what awaited friends who stayed in Cameroon, struggling to find decent jobs despite their education.
Throughout all of this, I was angry, very angry. I felt angry and ashamed when my parents couldn’t pay the annual public school fees of roughly US$16 on time, leading to my expulsion. Then angry at myself because I could not envision a way to support my family. Then angry at the ‘system’ that I felt was working against people like me. Then again angry at myself for being poor.
Looking back, I can see that anger, paired with an acute allergy to injustice, was my motivation – but agency is what energized me to take action. Moving to a big town, starting a hustle in last-mile delivery before last-mile delivery was a thing, moving to a foreign country, and so on. Years later, I realized that agency was my superpower – a capacity to make sense of my own environment and act upon that sense-making, even in the midst of uncertainty.
That brought me to Kabakoo Academies, the organization I co-founded in Bamako, Mali. We design and implement upskilling pathways for productive livelihoods in a region where less than 10 percent of youth have a shot at a formal job. The rest must face the precariousness of informally dominated economies.
We now work with over 40,000 learners, and what we’ve learned from our program data challenges a core assumption about youth programming in Africa: the binding constraint is not financial. It’s psychological.
Development’s “psych wave”
In early 2026, we went through thousands of video diaries recorded by young Malians as part of Kabakoo’s “Visualize to Succeed” module. For three minutes, each youth spoke to their phone camera about what stands between them and their goals. This wasn’t a survey. Nobody circled a number on a Likert scale. Each response was a minutes-long act of introspection as part of their learning journey. But when we coded every obstacle, a clear pattern emerged.
Psychosocial barriers like self-doubt, fear of failure, procrastination, or fear of judgment were named in 68 percent of the video diaries, while financial barriers were mentioned in 66 percent. The two are nearly tied in frequency, but radically different in depth. In the videos, financial obstacles get a single sentence. Psychosocial ones are wrestled with across entire paragraphs. Learners pause, circle back, contradict themselves, sometimes break down.
And when we tracked who they turn to for help, zero mentioned the state. Zero mentioned religious institutions. In a deeply religious society going through political transition, these young people articulate their futures without reference to either. One young woman put it plainly: “Moi, vraiment, j’ai confiance en personne. Si je forme une équipe, est-ce que quelqu’un ne va pas essayer de me nuire? Est-ce que quelqu’un ne va pas me voler? Est-ce qu’il m’abandonnerait?” [I truly trust no one. If I form a team, won’t someone try to harm me? Won’t someone steal from me? Would they abandon me?] We learned that youth rely on personal networks, self-made strategies, and their own agency.
Yet the broader livelihood sector has built on almost none of these. We had the microfinance wave; the results are out, and they’re sobering. Six randomized evaluations across four continents found modestly positive but not transformative effects, with no significant impact on household income. We might be right in the middle of the entrepreneurship wave; those results are pouring out, and they’re sobering too: more than a billion dollars spent annually, with average effects of under five percent on sales and a persistent tendency to confuse survivalist self-employment with transformational entrepreneurship. It’s just hard to get poor folks out of poverty, especially when they are living together with other poor folks in a poor country. Paths toward improved livelihoods are messy and involve more intricate challenges than distributing nets to reduce malaria-related deaths.
Now, I believe we are at the beginning of what I would call the psych wave, marking a shift from prior approaches toward mindset-oriented interventions for shared prosperity.
In 2025, Catherine Thomas and colleagues published a striking experiment. In rural Niger, they tested two 30-minute psychological interventions with women in extreme poverty. One was built from Western motivational science (i.e., personal goals, individual initiative). The other was designed from scratch using the women’s own mental models, which emphasized social harmony and collective advancement. After 12 months, the Western version moved self-efficacy scores but produced zero economic effect. The culturally grounded version significantly improved economic outcomes, including a 5.7 percentage-point drop in severe food insecurity – all from a single brief session a year before.
In Bambara, there is a concept that captures this. “Máa ka maya ka ca a yɛrɛ kɔnɔ.” It means, roughly, “the persons of the person are multiple in the person.” Similarly, every young person who recorded a video diary for Kabakoo carried both the obstacle and the response to it, in the same three minutes. The deficit and the agency coexist. Development programs that see only the deficit (e.g., lack of skills or capital) are missing half of what’s actually there. A whole half of what’s possible.
Working at this new frontier
At Kabakoo, our learning experience doesn’t start with coding lessons or basic bookkeeping. It starts with that visualization module – built on neuroscience and social psychology – where learners name their obstacles, list solutions, and add specifics like who can help, what resources they’ll need, and what exactly they’ll do.
Linguistic evidence from the video diaries shows that this approach works. For every deficit construction (“manque de confiance” [lack of confidence], “pas d’avenir” [no future]), learners produce nearly four agency constructions (“je vais” [I will], “je commence par” [I’ll start by]). As one learner, a young pastry entrepreneur, put it, speaking to his phone camera from Bamako: “Anticipez les obstacles, c’est comme prévoir les grumeaux dans la pâte. Tu sais qu’ils vont arriver, mais tu te prépares déjà à les lisser.” [Anticipating obstacles is like expecting lumps in the batter. You know they’re coming, but you’re already preparing to smooth them out.]
Because we are naive believers in the power of open learning, we decided to run a randomized controlled trial while we were even smaller, operating on a budget below US$300,000. Allies warned us. They worried funders might nail us down to whatever negative results came out. Their concern was valid. But we did it anyway. And what might seem like a deliberate act of self-sabotage turned out to be one of our best decisions.
We were able to instill a deep learning culture at Kabakoo, at a point where the organization’s core DNA was still in formation. We also got some exciting findings, despite a modest sample size shaped by our operating realities. At 36 months, for instance, we found that the income gap between our treatment and control groups had widened, not faded. This meant that every dollar invested in our learners generated US$5.60 in net economic benefit. And when we looked at the sequence, the pattern was clear: growth mindset shifted first; economic behavior followed.
Of course, addressing these inner obstacles only makes sense if you also face economic realities honestly. But new evidence says the development sector has that reality backward, too. In their 2025 paper “How Poverty Fell”, Paul Niehaus and colleagues track how households actually escape poverty across five countries. They find that in lower-income economies, people don’t escape poverty by switching to formal wage work. They escape by becoming more productive in their existing self-employed activities. The structural transformation story – leave the farm, enter the factory – doesn’t match what happens at the household level. Meanwhile, Bruno Crépon and colleagues’ evaluation of apprenticeships in Côte d’Ivoire found that even a well-designed, state-sponsored program raised earnings entirely through gains in self-employment, with no effect on access to wage jobs.
As Joel Bothello and Tim Weiss argue in their recent essay, Africa’s urban informal economies are not chaotic remnants waiting for formalization. They operate through sophisticated social orders that deserve investment on their own terms. So why do we keep training young Africans for formal jobs that the math says aren’t coming?
On top of that, technology is failing us. Literally. A 2025 study from the MIT Media Lab found that LLMs give answers that are 26.3 percentage points less accurate to users with lower levels of education and non-US origins. In other words, the people who might need AI-powered learning tools the most are getting the worst service.
We know this firsthand. At Kabakoo, we build AI in Bambara, deployed via WhatsApp, designed for US$50 phones with patchy data. We call our approach “highdigenous,” the intentional blend of high-tech and indigenous knowledges. One example: when we experimentally tested influencer marketing for our annual Bamako.ai festival, influencers with a combined 10 million followers produced just 300 registrations, while our grassroots outreach brought in almost 4,000. A trust survey of 1,079 Malian youth confirmed that social media influencers rank dead last in trust for career decisions; parents and teachers rank highest. Trust is local. Technology that ignores this isn’t helping anyone.
Lessons from our experience
Three things follow from this, if you design or fund youth jobs and productivity in Sub-Saharan Africa.
Address the inner obstacle directly. Put mindset-oriented interventions before or alongside technical training. The evidence from the Niger study highlighted above shows that culturally grounded approaches outperform imported ones. Our own study’s follow-up showed that mindset shifts precede and enable economic gains, even three years after the module. But too few fund this as a primary strategy.
Stop treating informality as the problem. The 90 percent of young West Africans who work beyond formal employment might be a policy failure. Sure. But they are certainly the economy. Fund their productivity, their agency, their social infrastructure; not a fantasy of mass formalization that a growing body of evidence says isn’t how poverty actually falls.
Build technology that speaks their language. If your EdTech platform doesn’t function in Bambara or Hausa or Wolof, on a low-budget phone, over WhatsApp, you’re building for a different continent.
In January 2026, the University of Lomé in Togo – one of West Africa’s largest public universities, with more than 74,000 students – formally adopted Kabakoo’s mindset-oriented modules for its graduates. The following month, Mali’s Ministry of Digital Economy signed a partnership with us to deploy Kabakoo programs in all public schools across the country. Governments in West Africa are already acting on what the research now confirms. You have to start where people actually are. Not as a feel-good trope, but really. In their minds. In their economies. And yes, in their social institutions and languages.
The evidence is there. What’s missing are more funders joining the party.
Kabakoo means “to wonder” in Bambara. We will keep wondering.
Yanick Kemayou is the Co-Founder and Chief Learning Officer of Kabakoo Academies.