Technical Assessment Platforms

eSkill launches "Linea," an AI layer that turns job descriptions into draft assessments

Source: eSkill · Aug 26, 2026

eSkill has been building validated pre-hire assessments for two decades, and its new AI layer, Linea, is squarely about making that twenty-year library faster to use rather than replacing the underlying assessment science. Launched August 26, Linea has two functions: Linea Build, which takes a job title or description and drafts a role-specific assessment in seconds, and Linea Search, which uses semantic matching to surface relevant questions from eSkill's library even when a recruiter's search terms don't line up exactly with how the library is indexed.

The system draws exclusively from eSkill's existing bank of more than 70,000 expert-validated questions, each tagged against ESCO occupational standards. That's a deliberate constraint: Linea isn't generating novel assessment content from scratch, it's routing and assembling from a pre-vetted, validated library. eSkill CEO Henry Tran framed the launch as making two decades of assessment work "surface through AI that works for the recruiter, not around them" — the emphasis being on acceleration and discovery, not autonomous decision-making.

A few details are worth noting for how eSkill is positioning trust and data handling. Linea runs on Anthropic's Claude models, no candidate personal data passes through the Linea layer at all, and recruiter/user submissions aren't used to train underlying AI models. Recruiters retain full editing control over any generated assessment and can inspect Linea's reasoning for why it recommended a particular question — a transparency feature that's becoming close to table stakes as AI moves deeper into hiring workflows that face increasing regulatory scrutiny. The feature is available immediately to existing eSkill customers, with no separate rollout phase mentioned.

For our tracking of the technical-assessment competitive landscape, Linea is a good example of the current wave of AI features in this space: assistive tooling layered onto an existing validated content library, rather than a wholesale AI-generated assessment product. eSkill is choosing speed-to-assessment and content discovery as its AI wedge, keeping the actual validated question bank as the trust anchor. It's a conservative, defensible move for an incumbent with two decades of validated content to protect, and it sets a bar other assessment platforms will likely need to match: AI-accelerated assessment creation, built on top of content whose validity long predates the AI layer.

Key Points

• Linea Build: turns a job title/description into a draft assessment in seconds
• Linea Search: semantic search across eSkill's question library, even when terminology doesn't match exactly
• Draws from 70,000+ expert-validated questions tagged against ESCO standards
• Built on Anthropic's Claude models; no candidate data passes through Linea and submissions aren't used for model training
• Recruiters keep full editing control and can review the AI's reasoning for each recommendation