AIQ Framework Reference

Complete documentation for the SCOREs dimensions, scoring weights, and validation matrix.

Framework Overview

The Universal AIQ Framework provides a standardized 0-100 score for individual AI competency, built on five dimensions that form the SCOREs acronym:

S - Study
Information
& Fluency
C - Copy
Evaluation
& Rigor
O - Output
Deployment
& Impact
R - Research
Innovation
& Contribution
Es - Ethical security
Safety
& Responsibility

Score Bands

Score Level What It Means
0-20UnawareNo meaningful AI adoption. At risk of displacement.
21-40UserBasic AI usage. Follows instructions. Needs supervision.
41-60PractitionerDaily productive use. Can evaluate quality. ~25% efficiency gain.
61-80BuilderDeploys reliable systems. Creates measurable business value.
81-95ArchitectAdvances practices. Mentors others. Trusted for critical work.
96-100PioneerIndustry-recognized contribution. Shapes how AI is used.

Evidence Levels

Level Multiplier Description
Level 1: Self0.70xSelf-report only. No validation.
Level 2: Peer0.85xPeer or manager confirmed the work.
Level 3: Verified1.0xAutomated logs or full audit with evidence.

Scoring Weight Tables

Your final score is calculated by weighting each SCOREs dimension based on three factors:

  1. Role — which skills matter most for your job function
  2. Company Type — organizational priorities that modify role weights
  3. Assessment Level — evidence confidence (you control this by getting peer validation or verified evidence)

Dual Scoring Weights

The framework uses two separate scores to distinguish what you know from what your organization has enabled:

Personal Readiness PRIMARY
"Am I ready to deliver AI value?"
Emphasizes Study + Copy (knowledge & evaluation skills)
Used for individual dashboards and comparisons.
Corporate Impact
"Has my org enabled AI delivery?"
Emphasizes Output + Ethical (deployment & governance)
Reveals organizational enablement gaps.

Personal Readiness Weights

Role Study Copy Output Research Ethical

Corporate Impact Weights

Role Study Copy Output Research Ethical
Gap Analysis (Personal − Corporate):
  • Positive gap (Personal > Corporate): You have skills your org isn't utilizing → advocate for AI projects
  • Negative gap (Corporate > Personal): Org is deploying faster than you're learning → invest in upskilling
  • Balanced (within ±10): Skills match opportunities → continue current trajectory

Company Type Modifiers

Multipliers applied to role weights, then renormalized to 100%.

Type Study Copy Output Research Ethical Philosophy
Startup 1.0x 0.7x 1.4x 1.2x 0.7x Ship it, learn, iterate
Enterprise 1.0x 1.2x 0.85x 1.0x 1.0x Reliable, scalable, governed
Aspirational 0.85x 0.85x 1.0x 1.2x 1.2x Build AI the right way

Validation Matrix

What evidence validates each dimension at each assessment level.

Dimension Level 1 (Self) Level 2 (Peer) Level 3 (Verified)
S - Study Self-reported sources Peer confirms knowledge Newsletter subs, course certs, reading logs
C - Copy Self-reported methods Peer reviews test cases Eval scripts, benchmark results, CI logs
O - Output Self-reported projects Peer confirms usage Git commits, deploy logs, usage metrics
R - Research Self-reported contributions Peer confirms novelty Publications, patents, model weights
Es - Ethical security Self-reported practices Peer confirms safety Audit logs, compliance records, training certs

Full SCOREs Rubrics

Detailed level descriptions for each dimension (Levels 0-5).

S - Study (Information & Fluency)

Where do you learn about AI? Can you explain why things work or fail?

L0No AI awareness. Avoids or fears the technology.
L1Mainstream news only. Passive consumption of hype/fear cycles.
L2LinkedIn influencers, YouTube "Top 10 Tools" content.
L3Developer blogs, release notes, AI-focused newsletters.
L4Technical reports, GitHub repos. Can explain why models fail.
L5ArXiv papers, model weights, source code. Predicts capability shifts.

C - Copy (Evaluation & Rigor)

How do you know if AI output is good? Can you prove it?

L0No validation. Blind trust: "It looks right to me."
L1"Vibes check." Runs prompt once, manually reviews.
L2Maintains test cases. Systematic manual Pass/Fail grading.
L3A/B tests models. Comparative benchmarks. Uses eval tools.
L4Automated evals. LLM-as-Judge. Quantified metrics (precision, recall).
L5Statistical confidence intervals. CI/CD for prompts. Regression testing.

O - Output (Deployment & Impact)

What have you built that others actually use? What value did it create?

L0Chat interface only. No deployment or workflow integration.
L1Personal productivity (Copilot, ChatGPT Plus). Time savings only.
L2Simple wrapper apps. Basic API integration. ROI not yet quantified.
L3Internal tools used by team. RAG pipelines. $10k+ verified savings.
L4Production agentic systems. Revenue-generating. External users.
L5Vertical AI platform. Fine-tuned models. $100k+ verified value.

R - Research (Innovation & Contribution)

Do you advance the field or just consume it?

L0Treats AI as magic. No understanding of mechanisms.
L1Conceptual understanding: tokens, temperature, context windows.
L2Architectural knowledge. Understands Transformers. Implements papers.
L3Contributes: fine-tunes models, publishes weights, shares methods.
L4Researches: novel architectures, publishes at conferences.
L5Invents: paradigm-shifting discoveries. Industry-recognized impact.

Es - Ethical security (Safety & Responsibility)

Can you be trusted with AI? Do you use it safely?

L0Dangerous. Pastes PII into public models. Ignores bias.
L1Compliant. Follows rules. Uses only sanctioned tools.
L2Cautious. Fact-checks outputs. Human-in-the-loop for decisions.
L3Proactive. Tests for hallucinations. Documents failure modes.
L4Guardian. Catches risks in others' work. Designs safety protocols.
L5Leader. Shapes org policies. Trains others on safe practices.

For Administrators

Share pre-configured assessment links with your team to ensure everyone uses the same role and company settings.

Pre-configured Assessment Links

You can generate assessment URLs with settings pre-filled. When users open these links, their role, company type, and scoring distribution will be automatically selected.

Available URL Parameters

Parameter Values Default Description
role General, Developer, Researcher, Support, Leader General Sets the primary role for dimension weighting
companyType Startup, Enterprise, Aspirational (none) Applies company-specific weight modifiers
distribution bellCurve, linear, progressive, sigmoid bellCurve Point distribution curve (Advanced Options)

Example URLs

Developer at a Startup:

https://sagearbor.github.io/ai-skill-eval-kit/level1.html?role=Developer&companyType=Startup

Leader at an Enterprise company:

https://sagearbor.github.io/ai-skill-eval-kit/level1.html?role=Leader&companyType=Enterprise

Researcher with progressive point distribution:

https://sagearbor.github.io/ai-skill-eval-kit/level1.html?role=Researcher&distribution=progressive

Using the Share Settings Button

On the assessment page, after selecting your desired settings:

  1. Configure the role, company type, and distribution as needed
  2. Click the "Share Settings" button (located above the assessment form)
  3. The URL with your current settings is copied to your clipboard
  4. Share this link with your team via email, Slack, or any messaging platform

Tip: Only non-default values are included in the URL to keep it clean. If you select "General" role with no company type and "bellCurve" distribution, the URL will be the plain assessment page.

Use Cases

Take the Assessment →