Signal › Contributors

Amara Diallo

I spent three years at McKinsey in the Education practice, from 2015 to 2018. My clients were universities, state education departments, and large ed-tech companies. The work was interesting in the way that consulting is always interesting: you learn a lot about a lot of industries, and you produce beautifully formatted slide decks that may or may not change anything.

What McKinsey taught me was that education is a market where the incentives are catastrophically misaligned. Universities optimize for rankings and research output, not student outcomes. Employers say they want skills-based hiring but still filter for pedigree. Ed-tech companies optimize for engagement metrics that don't correlate with learning. Everyone talks about "closing the skills gap" while actively maintaining the structures that created it.

I left McKinsey to join Coursera in 2018, eventually becoming VP of Product for the enterprise business (Coursera for Business). My team built the product that companies use to upskill their workforce: the course recommendation engine, the skills assessment tools, the analytics dashboard that L&D leaders use to justify their budgets to CFOs. We grew the enterprise business from $30M to $140M ARR during my tenure.

The most important thing I learned at Coursera was the completion rate problem. The average MOOC completion rate is about 5%. For enterprise customers paying $400 per seat, that number is slightly better, maybe 15-20%. Nobody in the industry talks about this honestly because the business model doesn't depend on people finishing courses. It depends on people starting them. Or more precisely, it depends on L&D teams buying seats. Whether anyone learns anything is, economically speaking, a secondary concern.

That realization is what made me leave and start writing. I believe technology can transform education and workforce development, but only if we're honest about what's working and what isn't. Most ed-tech coverage is either breathless optimism or cynical dismissal. I try to write with the specificity of someone who's seen the data and the nuance of someone who understands why the incentives are broken.

I'm based in London now. I grew up in Dakar, studied in Paris, and worked in New York before moving to the UK. I speak French, Wolof, and English. I mentor first-generation college students on weekends and I'm working on a book about the economics of credentialing.

Experience

Articles by Amara Diallo (4)

The AEO Case Study: How to Structure Client Stories That AI Assistants Actually CiteMost case studies are written for human buyers at the bottom of the funnel. The AI-citation-optimized case study is a different document with a differ · May 25, 2026AI-Generated UGC: Detection, Penalties, and the New AEO Quality Bar in 2026Claude Skills lets vendors publish installable capabilities Claude can call directly. Stripe, Linear, and Notion are early movers, and the skill listi · May 25, 2026Multimodal Search Optimization: Image, Audio, and Text AEO in the Same PipelineApple Intelligence, Google Gemini Nano, and Qualcomm AI Hub pushed inference onto smartphones — and into compliance-locked contexts like K-12 classroo · May 25, 2026Local-Language LLMs Are Eating Emerging Markets. AEO Strategy Splits.Towards Data Science gets cited in 41% of ChatGPT data-science answers. The Startup almost never does. We profiled 14 Medium publications, ran 1,800 L · May 26, 2026