Skillwitness

Employers want to understand whether candidates can use AI tools responsibly, but self-report and simulated quizzes reveal only part of a workflow. Skillwitness can run a bounded, job-related work sample and produce attributable evidence about planning, verification, iteration and final deliverable review. The supplied research confirms several live assessment products, including one that captures AI interactions during take-home coding work, so the category and a close concept already exist. The unverified delta is a portable evidence artifact and carefully governed workflow data—not an empty market. Continuous employee surveillance and a universal productivity profile should not be the default product: tool-use frequency, prompt count, suggestion acceptance and commit patterns do not measure effort, competence, value or output quality. Employment decisions require job relevance, notice, lawful basis, accommodations, accessibility, retention limits, adverse-impact testing, human review and appeal. A signed artifact protects provenance; it does not make a rubric valid or a candidate competent. Consent, collection, task evidence, rubric observation, assessor disposition, candidate review, hiring authority, employment outcome and later performance remain separate. Success is a more structured and contestable work-sample review—not anti-cheat certainty, proof of being AI-native or continuous ranking of workers.

A talent-acquisition or technical-hiring team with counsel and people-operations oversight for a specific job-related work-sample process.