From Principal to Founder: Adeel Khan on Building AI for a Million Teachers
Key points
Key takeaways from a Verticals interview with MagicSchool Founder & CEO Adeel Khan (June 2026):
Teacher to founder. Khan taught in Atlanta public schools, became an assistant principal in Denver and founded Conservatory Green High School — the top-performing public high school in Denver Public Schools, with 100% of graduates accepted to four-year colleges — before a 2022 sabbatical ended at the ChatGPT moment. With no software background, he launched the MagicSchool MVP in May 2023.
PLG explosion. MagicSchool went from zero to 1M educators in five months, entirely organic growth, and remains the top search result for "AI for teachers"; districts representing roughly one in five children in the US now partner with the company, which is used in nearly every school building in the country and across 160 countries.
Free-first monetisation. The product stayed completely free until January 2024, when the company started monetising around the 1M-user mark; Khan's stated heuristic was that the most impactful companies give the product away and figure out the business later.
$1,000 pilots. To convert pent-up demand into contracted revenue, MagicSchool signed ultra-low-cost spring pilots — $1,000 per school site and $2,000 per district regardless of size — with the full ticket price contracted for the July renewal cycle; every pilot renewed at full price, vindicating the bet ahead of the Series A.
Sales evolution. The first year of selling was pure order-taking on inbound demand, then a K-12-experienced CRO assembled a professional sales team and the motion shifted to strategic account executives, though the company still enjoys strong market pull and referenceable districts.
Domain expertise as the moat. Roughly 75% of the early team had classroom experience, and today an estimated 85–90% of go-to-market hires are former educators — including solutions architects recruited out of district technology roles — reflecting the belief that domain knowledge is harder to train than sales skill.
GPT wrappers got more valuable. Khan argues application-layer companies were underpriced: teachers spend one to two hours a day on computers and will not assemble Claude skills and connectors, so a purpose-built product carrying user context — school, students, performance data — beats horizontal models; he cites Jasper's bleed to ChatGPT as the counter-case that did not repeat in education.
Copycats don't show up in deals. Despite roughly a hundred lookalike products from curriculum companies and systems of record such as PowerSchool, competitors rarely appear in MagicSchool's deals; some legacy players are moving "headless" in the Salesforce mould, opening partnership paths, while brand and relationship remain the durable moat.
Read more: Verticals — YouTube
