How scoring works

AI Fluency Standard.

An evidence-based assessment of how employees use AI. Four dimensions combine into one score out of 800, showing where each person is today and where to focus next.

Create company account CAD $100 per employee. No subscription required.

Four bands. A practical next step.

0–800 points
What this looks like at work

Emerging

Occasional personal use, with no verification habits.

People reach for ChatGPT or Copilot for first drafts and summaries on their own initiative. No company data, no repeatable process, and no check on the output — quality depends entirely on the individual.

Next step

Build verification habits

Choose a repeatable task. Check the output against trusted sources and record where human review is needed.

The right target depends on the role. Not every employee needs to reach Advanced.

See the scoring methodology
The Model

AI capability. A common standard.

Our proprietary scoring model turns evidence of AI capability into a number leadership can act on. Built in-house, developed in higher education, and now applied to company teams.

Then / Credit risk

From a judgment call
to a common score.

Credit scores gave lenders a shared way to assess risk, beyond individual relationships and subjective judgment.

One standard. Decisions at scale.
Now / AI capability

The same idea.
A new kind of capability.

The Fluency Score brings a common measure to AI capability. Compare functions, identify gaps, and know where to invest.

One score. Four dimensions. A clearer next move.

A purpose-built assessment, not a self-scored quiz.

Inside the 800-point score
Scoring Methodology

One score. The full picture.

Four equally weighted dimensions. 200 points each. A composite out of 800 that shows both the headline result and the capability behind it.

AI Fluency Score

Illustrative result

720/ 800

Advanced

0800

The headline tells you where someone stands. The breakdown tells you why.

Behind the number

4 dimensions

Technical AI Depth

188 / 200

How well someone operates AI tools, prompts, models, and controls.

Applied AI Fit

176 / 200

How effectively AI improves real work within someone's function.

Risk–Reward Competency

164 / 200

Judgment on hallucination, bias, privacy, and when not to trust AI.

AI Mini-Case Analysis

192 / 200

Decision quality in open-ended business scenarios, beyond multiple choice.

See where your team stands.CAD $100 per employee. One assessment.

Create company account
From evidence to action

A score you
can trace.

The portfolio-evidence approach developed in higher education, applied function by function. Follow the work from the first submission to the investment decision.

Illustrative outputs
  1. Collect evidence

    Real work product, from documents to workflows and shipped tools. Evidence differs by function; the scoring rubric stays consistent.

    Evidence packet By function
    • AI-assisted campaign briefMarketing
    • Documented team workflowOperations
    • Shipped internal toolEngineering
  2. Score each dimension

    Responses are auto-graded against four 200-point rubrics, calibrated to human expert graders. The four sub-scores combine into one result out of 800.

    Individual result Four dimensions
    720 / 800Advanced
    Technical AI Depth188 / 200
    Applied AI Fit176 / 200
    Risk–Reward Competency164 / 200
    AI Mini-Case Analysis192 / 200
  3. See the distribution

    Individual scores roll up by function, exposing the range of capability that a company average can hide.

    Function distribution Operations
    416 / 800Function average
    240320400480640
    0800

    Five sample scores. The average alone hides the spread.

  4. Identify the gaps

    Compare each function with its target. The gap map shows where to invest; the sequenced roadmap sets out what closing those gaps requires.

    Investment priorities Points to target
    FunctionNowTargetGap
    Operations416600184
    Marketing464600136
    Engineering63270068

    Operations has the largest gap in this example. Targets depend on the role and business need.

How points and rubrics work
Four sub-scores, 200 points each
Every employee is scored across the four dimensions above. The four sum to one composite out of 800, so leadership sees both the headline number and the dimension driving it.
Scenario-weighted items
Points reward decision quality, not recall of tool names. Harder scenarios carry more points, and difficulty escalates until the employee's ceiling is found.
Faculty-calibrated rubric
Open-ended case responses are graded against a rubric calibrated by Ivey / DAN Management faculty, keeping scores comparable across employees and companies.

Start with your team's evidence.
Know where to act next.

Create company account
Validation & Rigor

Built to be checked.

Adaptive assessment. Expert-calibrated grading. Ongoing reliability checks.

Where validation standsA published academic validation study is a near-term objective, not a result we claim today.

Adaptive item selection

The computerized-adaptive assessment adjusts the difficulty of the next item after each answer. Scenario-weighted items reward decision quality, not recall of tool names, and place employees on the 0–800 scale in a sitting of up to 60 minutes.

Human-calibrated, LLM-assisted grading

Open-ended scenario responses and portfolio evidence are scored against the sub-score rubrics using LLM-assisted grading, calibrated against human expert graders from our university deployments. Human-graded evidence is the ground truth the model is checked against. The rubric is calibrated by Ivey / DAN Management faculty.

Reliability and validation

We track agreement between machine and calibrated human graders, test-retest reliability across quarterly re-scores, and item-level difficulty and discrimination as the response corpus grows. A published study with an academic partner, covering inter-rater reliability coefficients and predictive validity against job performance, is a near-term objective.

AI Fluency Standard™. Trademark filed in Canada; international filing in progress.

What you get

Your board-ready
AI roadmap.

One report connects the investment decision to the work ahead: where each function stands, where it needs to go, and what to do next.

AI capability roadmap

Report previewIllustrative data, targets, and actions

Current score distribution

Evidence-scored capability by function, on the 0–800 scale.

Engineering
0500690 current range800
Target bandAdvanced601–800
Marketing
0280480 current range800
Target bandProficient401–600
Operations
0160370 current range800
Target bandDeveloping201–400

Targets reflect the role and the business need. Advanced is not the goal for every function.

Sequenced closing plan

Example priorities and accountable teams

  1. 01
    Align the tools

    Select approved tools and define where they fit in each team's work.

    Function leads + IT
  2. 02
    Close priority skill gaps

    Focus practice and training on the capabilities furthest from the target band.

    Function leads + L&D
  3. 03
    Put controls in place

    Set review requirements, data-handling rules, and accountability for AI-supported work.

    Risk + governance
Sophrosyne SystemsAI Fluency Standard

Support beyond the report.

Turn the roadmap into a learning rhythm with monthly scenarios, answer explanations, company-level insight, and verified reassessment.

Explore learning plans
How It Runs

From baseline to board-ready metrics in days.

Four steps that connect an official baseline with targeted learning, practice, and verified progress.

01

Verified baseline

Employees take the role-specific Verified AI Fluency Assessment. Each official result is reported as a Verified AI Fluency Score out of 800.

02

Company-wide view

Individual scores roll up into the Company Progress dashboard, showing strengths, gaps, and function-level patterns.

03

Targeted learning

Monthly scenarios and answer explanations help each employee build the capabilities their baseline identified.

04

Practice and verify

Practice Reviews in months 3 and 6 support learning, followed by a Verified Reassessment in month 6 to measure progress.

Company Progress
Illustrative Preview

The dashboard behind the learning.

The Company Progress Plan brings your teams' scores, the 0–800 distribution across bands, practice cadence, and month-six reassessment into one view — connecting what people are learning with the progress leadership needs to see.

platform.sophrosynesystems.org

Analytics

Company

Sample Company

FY 2026

6 Teams

Data Sources

Slack
HRIS
Google Workspace

This is what your dashboard looks like once your team is onboarded.

Create company account See how scoring works

Measure. Learn. Verify.

Know where your people stand.
Help them move forward.

Start with a verified baseline. Build capability with purpose. Return to the score and see what changed.

Create company account

CAD $100 per employee / No subscription required