The 7 meta-capabilities laituri.ai measures: structural thinking abilities that transfer across roles, industries, and AI tools.
What We Measure
We don't test "Do you know Python?" We measure structural thinking abilities that transfer across every role, industry, and AI tool.
Tool-specific tests are obsolete in months. A participant who "knows ChatGPT" today may be using a completely different AI stack next year. We measure the thinking patterns underneath: the capabilities that let someone learn any new AI tool fast.
A meta-capability is a structural cognitive pattern, like how a person breaks down ambiguity, filters noise, or improves through iteration. These patterns predict AI performance across tools, roles, and industries.
What the assessor measures
Breaking complex, ambiguous challenges into structured, workable sub-tasks.
Can the participant identify the real problem inside a messy prompt? Do they separate symptoms from root causes? Do they create a logical sequence of steps before acting? This is the foundation of effective AI collaboration.
Connecting multiple steps, agents, or outputs into a coherent end-to-end solution.
Does the participant think in pipelines, not just single prompts? Can they anticipate how one AI output will affect the next step? Can they design a multi-stage workflow rather than a one-shot response?
Using AI tools precisely and efficiently to produce high-quality work output.
Does the participant write clear, context-rich prompts? Do they provide the right constraints and examples? Can they get a useful result from an AI tool without excessive trial and error?
Working alongside AI as a genuine thinking partner, not just a text generator.
Does the participant treat AI as a collaborator? Do they verify outputs critically? Can they combine their own domain expertise with AI output to produce something better than either could alone?
Pivoting quickly and confidently when conditions change or AI output is wrong.
What does the participant do when the AI gives a bad answer? Do they freeze, accept it, or reframe the problem? How quickly can they adjust their approach without losing momentum?
Identifying what matters in large, complex, or noisy AI-generated outputs.
Can the participant find the key insight in a 2,000-word AI summary? Do they know which parts to trust, which to verify, and which to discard? Can they synthesize across multiple sources without losing the thread?
Refining outputs through structured, purposeful prompting loops.
Does the participant stop at the first response, or do they improve it systematically? Can they diagnose why an output is weak and apply a targeted correction? Do they know when to stop iterating?
Every response is evaluated by AI on all 7 dimensions simultaneously. Scoring is consistent, objective, and immediate: no human rater variation.
All 7 scores combine into a single Composite Score (0–100). HR sees the breakdown as a radar chart and the composite for quick ranking.
Prompt Skills and Answer Quality count 50/50 for every role, and all 7 capabilities feed those two halves evenly. Fixed equal weights predict at least as well as tuned ones, so there are no weightings to set or to get wrong. Role-specific must-haves are handled by the eligibility checklist instead.
Start a free assessment or explore the HR dashboard.