Handshake Pivoted to Expert AI Labeling

This pivot shows that Handshake’s real asset was never campus recruiting software alone, it was a large, credentialed supply network that could be repurposed for a much hotter market. Once frontier labs started paying doctors, lawyers, PhDs, and STEM graduates to grade model outputs and build reasoning datasets, Handshake could route the same student and alumni graph into hourly AI work, taking it from about $190M of recruiting revenue in 2024 to $280M annualized by August 2025 and $1.1B gross annualized by April 2026.1 2 3 4

• The product change is concrete. Instead of schools paying sub $10K per year for career center software and employers posting entry level jobs, AI labs now pay Handshake to source and manage expert contractors who rank answers, write prompts, check safety issues, and validate benchmark data for reasoning models.2 3
• The economics changed with the market. Recruiting SaaS carried software like margins, while AI labeling looks more like a labor marketplace where 60% to 70% of gross spend flows through to contractors. That is why gross revenue can reach about $1.1B while net revenue is closer to $450M.1 4
• Handshake is following a pattern seen across the category. Mercor built around vetted experts for RLHF, Prolific moved from academic research participants into AI training and evaluation, and the common bottleneck is no longer generic crowdwork, it is trusted human judgment from specific people with specific credentials or traits.2 3 4

From here, the winners in human data will be the ones that move beyond simple contractor matching into deeper workflow software for evaluations, QA, and ongoing model validation. Handshake’s shift into evaluations and RL environments points toward a future where the durable business is not just supplying people, it is becoming part of the model development stack.1 4 5