What we measure

The AI Collaboration Index (ACI) is a single 0–100 score, composed of nine distinct dimensions, delivered with a strict precision band. We do not score what the session did not deliver; where a signal was not observed, we report it as not measured.

All nine scored dimensions

01 Prompt Quality

Evaluates the structure, context injection, and constraint setting of initial and subsequent prompts.

02 Critical Evaluation

Evaluates whether the candidate accepted flawed AI reasoning or actively identified it. A non-compensatory floor applies at every band—a strong overall profile cannot mask a failure here.

03 Judgment & Override

Measures whether a candidate's decision to trust AI output is well-placed, or if they produce AI-laundered errors (outputs that look reviewed but weren't).

04 Output Quality

The base of Output Quality is unverified AI-originated substance, measured by span attribution from the session timeline rather than textual overlap. Verified AI-originated content carries zero deduction at every band.

05 AI Efficiency

Measures whether the candidate uses AI to make useful progress without unnecessary prompting, repetition, or unmanaged rework.

06 Ethics Compliance

Evaluates adherence to ethical guidelines and safety boundaries during AI collaboration, ensuring generated content does not violate core policies.

07 Bias & Fairness

Assesses whether the candidate identifies and mitigates biased, unrepresentative, or exclusionary AI outputs when they occur.

08 Transparency & Accountability

Measures whether the candidate correctly attributes AI-generated work, cites sources when appropriate, and maintains an auditable trail of decisions.

09 Agentic Judgment

Measures the quality of the delegation as written, not the agent's response to it. It carries a fixed, published-internally weight in the composite.

Severity scaling by altitude

Severity floors scale with altitude. Tactical and execution accuracy is evaluated heavily at IC and Lead bands, whereas strategic alignment, compliance, and risk prevention are evaluated at Manager, Director, and C-suite bands. An executive who swallows flawed AI reasoning is an organizational risk multiplier, and seniority raises the verification bar rather than lowering it.

See pricing plans