AI-Powered Reviews
Let AI explain what the scores mean
Assess4me uses a state‑of‑the‑art AI model to evaluate every candidate at three levels: per game, per category, and per vacancy. Each level produces skill‑by‑skill scores with written explanations so you always know why someone scored the way they did.
14-day free trial · Cancel anytime
The Science
Built on the Assess4me Integrated Model
AIM is the framework behind every score you see. It connects how candidates play to validated psychological constructs, so the numbers actually mean something in a hiring decision.
Evidence-Centered Design
Every game action is traced back to a specific competency through a documented evidence model. Scores are auditable, not assumed.
Stealth Assessment
Measurement runs in the background while the candidate plays. There is no questionnaire to game, which cuts social desirability bias and reveals real behavior.
Process Data, Not Just Outcomes
AIM looks at the order and timing of decisions. Two candidates can reach the same result through very different reasoning, and the report shows that.
Computational Psychometrics
Bayesian updating, item response theory, and behavioral pattern models work together so scores stay reliable across sessions and roles.
Neuroscience Grounded
Each competency maps to specific neural systems studied in cognitive neuroscience. We do not collapse different processes under one label when the brain treats them as separate.
Human in the Loop
AIM is decision support. The recruiter still makes the call. Reports include narrative context and interview probes, never an automated yes or no.
AIM draws on Evidence-Centered Design (Mislevy & Haertel), stealth assessment (Shute), computational psychometrics (von Davier), process mining (van der Aalst), and cognitive neuroscience research. The model is documented internally and maintained as the scientific basis for academic validation work.
Three-Level Pipeline
AI reviews at every depth
Instead of one flat score, the pipeline evaluates candidates at three nested levels so you can zoom in on a single game or zoom out to the full vacancy fit.
Level 1
Game Evaluation
Each game is evaluated individually. The AI receives the candidate's normalized score, time spent, and retry count, then produces a score (0–100) and written rationale for every skill that game tests.
Example: a memory game might yield memory 90/100 and risk_management 75/100, each with its own written explanation.
Level 2
Category Evaluation
All games within a category are aggregated. The AI reviews the combined results and scores each category-level skill, adjusting the pre-computed average by up to ±5 based on qualitative patterns it detects.
Categories group related games together, like 'Cognitive Speed' or 'Problem Solving', so you get a meaningful roll-up.
Level 3
Vacancy Evaluation
The final synthesis. AI scores every skill across the entire vacancy, combining all category and game results into one structured report with global skill scores, per-category breakdowns, and an overall explanation.
This is the report recruiters see: a single score, skill radar, and narrative for how well the candidate fits the role.
Built for Trust
Responsible AI by design
Recruitment decisions need to be defensible. Every part of the AI pipeline is built for transparency, fairness, and reproducibility.
Fully Explainable
Every skill score comes with a written rationale. No black-box numbers. Recruiters can read exactly why the AI scored each skill the way it did.
Deterministic & Fair
The AI is configured for maximum consistency. Identical performance always produces identical scores, with no variance between evaluations.
Fast Results
A full three-level vacancy evaluation typically completes in 1–3 seconds. Caching cuts redundant AI calls by 60–70%, keeping costs low.
Transparent Scoring
Raw scores are computed deterministically before AI interprets them. The AI cannot invent data. It only explains the numbers the system already calculated.
Structured Output
Results are returned as structured JSON: global skill scores, per-category breakdowns, and per-game details. Ready for dashboards, radar charts, and comparison views.
Smart Caching
Every AI evaluation is cached by input hash. If two candidates produce the same game metrics, the cached result is returned instantly, saving time and cost.
What You See
From AI scores to hiring decisions
The recruiter dashboard turns structured AI output into clear, actionable views so you spend time deciding, not deciphering.
Skill-by-Skill Scores
Every evaluated skill receives a normalized 0–100 score plus a written explanation. View them as progress bars, radar charts, or exportable data.
Written Rationales
The AI doesn't just give a number. It explains what it observed. Game metrics like time spent, retries, and raw score are referenced in every rationale.
Side-by-Side Comparison
Pick any two candidates and compare their full AI reviews, skill breakdowns, and scores in one view. A clear winner or tie indicator helps you decide faster.
Ranked Leaderboard
Candidates are ranked by their aggregated vacancy score, a weighted sum produced by the deterministic pipeline and refined by AI. Sort, filter, and drill into any result.
FAQs
Start assessing candidates the right way
Combine game-based soft skill assessments with hard skill quizzes and let AI help you find the best fit for every role.