One in five young people globally are not in employment, education, or training (NEET). In many developing countries, there simply aren’t enough jobs. And even where positions exist, a different problem emerges: Young people and employers can’t connect. Weak networks, geographic barriers, and lack of information create invisible walls between talent and opportunity.

Employment services can bridge this gap, especially for disadvantaged youth without strong connections. But in low- and middle-income countries facing the worst youth unemployment, these services, traditionally publicly provided, are often limited or non-existent.

The digital opportunity

 Digital job platforms are reimagining how people find work—aggregating thousands of vacancies, using algorithms to match talent with opportunity, and helping jobseekers showcase their skills to employers, with successful solutions adapted to locations with limited connectivity.

The results are promising. HaHuJobs in Ethiopia has significantly improved job finding rates by pulling together scattered vacancy data into a digital platform. In South Africa, the SAYouth platform operated by the NGO Harambee, supports over 4.2 million young people with digital job matching services.

Artificial intelligence (AI) is opening new frontiers. In Sweden, an AI uses jobseekers’ click history to generate personalized job recommendations, increasing employment within six months. Suggesting alternative occupations encouraged jobseekers to broaden their search for new job opportunities.

Other digital tools are focused on helping jobseekers showcase their skills more clearly to employers. In East Africa, Fuzu integrates short training modules—ranging from interview skills to digital literacy—into its job platform, so jobseekers can strengthen their profiles while they search.  Shomvob in Bangladesh uses AI to generate CVs from basic information, helping youth with limited digital skills compete more effectively. Related approaches, including skills assessment tools such as SkillCraft, are being explored in different contexts including in South Africa. Early pilots of conversational AI tools in Kenya and South Africa are showing promise in helping jobseekers showcase capabilities from informal and unpaid work experiences that traditional resume formats often miss.

These innovations can expand opportunities for jobseekers and firms at scale and at lower marginal cost than traditional public employment services. Their value is greatest when they plug into the wider labor market architecture and complement in-person services.

The hard truth: Digital isn’t automatically inclusive

But digital solutions can make inequality worse. In low- and middle-income countries women are 7% less likely to own a phone and 19% less likely to have access to mobile internet. For women already facing a massive labor force participation gap, digital-first approaches risk pushing them further to the margins. Add concerns about safety and privacy, and many women simply won’t engage.

Algorithmic bias is another risk. The widely cited case of Amazon’s AI recruiting tool, which penalized women’s résumés, illustrates the dangers of adopting digital tools without careful consideration of its risks.  Preventing bias requires high-quality, diverse datasets, and ongoing human oversight. The technology is only as good as the data behind it and the humans monitoring it.

Making digital work for everyone

So how do we harness the power of platforms while ensuring they don’t leave people behind?

First, governments need to uphold transparency while protecting privacy. How is data being used? Who controls personal information? Are algorithms making decisions alone, or is there human oversight? These aren’t nice-to-haves, they’re essential for adoption.

Second, meet people where they are. Not everyone has a smartphone or unlimited data. SAYouth’s WhatsApp integration works because that’s where young South Africans already communicate. In Ghana, the Text4Jobs initiative uses low-tech solutions so no smart phones are needed. Partnerships with telecom providers to lower data costs can expand reach even further.

Third is the need for credible skill certification systems. Many young people—especially women—have real skills that don’t fit traditional résumés. Managing a household budget is financial management. Coordinating family responsibilities is project coordination. Platforms can help translate these experiences into identified skills and capabilities that employers value. Harambee is pioneering this with the University of Oxford and Tabiya on SAYouth platform, offering important insights for how governments and partners use digital capability profiles and skills taxonomies beyond what formal education misses. This approach is grounded in local realities and proves especially powerful for reaching excluded youth.

Fourth, connect the dots. Right now, job platforms operate in silos. Private platforms don’t talk to NGO platforms, and neither connect with government systems. But when they do integrate, results improve dramatically. In Ethiopia, public employment services became more effective in helping jobseekers find job opportunities once a private provider shared their curated up-to-date vacancy data with them —especially for lower-skilled women, the most marginalized group.

Finally, think bigger. Link platform data to labor market forecasting to predict where jobs are heading and what skills are in rising demand. Use these insights to design skills development and employment programs aligned with current and future labor market needs, ensuring well-tailored solutions.

Execution matters: Institutionalizing digital labor market innovations

Digital tools are essential building blocks for smarter, more systematic labor market policy. When designed for inclusion and embedded in public systems, they can unlock opportunities at scale. The challenge now is disciplined execution: Scale what works, integrate platforms with public employment services, and keep safeguards and human-centered design at the core—so digital solutions widen doors, not gaps.



Source link