Linea Build is eSkill’s AI-assisted assessment builder. You provide a job title or job description, and Linea Build drafts a skills assessment for that role using questions from eSkill’s validated question library. This FAQ explains what the AI does, what it doesn’t do, and how we address the questions customers ask most about accuracy, fairness, privacy, and oversight.
No. By default, every question in a Linea Build assessment comes from eSkill’s existing library of approximately 60,000 professionally developed questions. The AI never writes, rewrites, or modifies question content. Its job is to understand your role requirements and select the most relevant questions from that library — the same questions available to any eSkill customer building an assessment by hand.
If you have created custom questions in your instance of eSkill using the Author feature, you have the option to grant access to these questions so that they may be used by Linea Build when relevant to a job description. Contact our support to enable access to your custom questions.
Linea Build uses AI for two things. First, it reads the job title or job description you provide and turns it into a structured job profile: the role, its context, and the skills, tools, and knowledge areas that matter. You see this profile and can edit it before anything else happens. Second, it searches the question library from multiple angles and selects a draft set of questions that fit the profile, balancing primary and secondary skills, tools, difficulty, and overall assessment length.
No. Linea Build is used only when an assessment is being created, before any candidate is involved. It does not score candidates, rank applicants, screen resumes, or recommend hiring decisions. Once an assessment is generated, it works exactly like any other eSkill assessment: candidates answer validated questions, and scoring follows eSkill’s standard methodology. Hiring decisions remain entirely with you.
No. Candidates take a standard eSkill assessment. The AI is only involved on the administrator’s side, during assessment creation.
Yes, at every stage. The job profile the AI extracts is shown to you for review and editing before any questions are selected. The generated assessment is a draft: you can add, remove, or replace any question before publishing. Linea Build also shows a short explanation for why each question was chosen, so you can evaluate its reasoning rather than take it on faith.
If you enter just a role name, Linea Build walks you through a short, guided builder, one step at a time — skills, tools, knowledge, and role context — suggesting options you can accept, edit, or replace with your own. Nothing enters the profile without your confirmation. This path is designed for people who aren’t sure what to assess and want to work it out step by step.
The selection quality comes primarily from the question library itself. Each question is tagged with structured labels for skill, role context, difficulty, and industry fit, using ESCO — the European Union’s standardized occupation and skills classification — as the backbone of skill labelling. Rather than matching keywords, Linea Build searches this enriched library from multiple semantic angles and balances the results, so every aspect of the role is represented, then applies assessment-design rules to make the final selection.
Yes. Linea Build draws exclusively from eSkill’s validated content library — the same professionally developed question bank that underpins all eSkill assessments. Using AI to select questions does not change the content, validity, or scoring of the questions themselves.
Only the information you provide to build the assessment: the job title or job description and the resulting job profile. No candidate data, candidate responses, or personal information is processed by the AI. Job descriptions are typically public documents by nature, but we recommend not pasting confidential internal information into the builder.
No. Linea Build uses commercial AI models accessed through enterprise API agreements, and inputs submitted through the platform are not used to train those models.
Linea Build uses commercial large language models from Anthropic (the Claude model family), accessed via API under enterprise terms. Different tasks use different models, chosen for the right balance of speed and reasoning quality.
Several design choices limit the ways bias could enter the process. The AI selects from a validated question library rather than generating content, so there is no free-form generated text in front of candidates. Skill labelling is based on ESCO, an externally maintained, standardized vocabulary, rather than an ad-hoc internal taxonomy. Every AI output is a draft subject to human review, and the AI plays no role in evaluating candidates.
Yes. Every generated assessment is logged: the questions selected, the stated reason for each selection, the source job profile, and a timestamp. This record can be retrieved after the fact, so you can always show how and why an assessment was composed.
Linea Build produces a draft that you review, edit, and approve — the assessment you publish is your selection, supported by AI. As with any assessment, we recommend reviewing generated assessments against your organization’s hiring policies and applicable regulations before use. Our team is happy to support compliance reviews by providing the audit records described above.
Linea Build is built with automatic fallbacks: if the primary AI model is slow or unavailable, a backup model completes the task. If needed, you can always build or edit assessments manually using eSkill’s standard tools — Linea Build is an accelerator, not a dependency.
Have a question we didn’t cover? Contact your eSkill account representative or support team — we’re glad to walk through how Linea Build works in as much detail as you need.
