Back to AI Assessment

    How scoring works

    The StackUp AI Maturity Assessment scores your organisation on a 0–100 scale across the four questions every executive asks. Here's exactly how the score is calculated, what each band means, and how the AI Risk Rating is derived.

    The 5-level rubric

    Every scoring question is rated against the same 5-level rubric. Level 3 is intentionally the most common reality — "we have it, but it's not really being used." Levels 4 and 5 always involve some combination of governance, frequency, accountability, and visibility.

    1
    Absent or negative
    2
    Informal / ad hoc
    3
    We have it but it's not really used
    4
    Documented and applied
    5
    Mature, governed, measured, continuous

    The four executive questions

    37 scoring questions grouped under the four questions every executive asks. Plus 6 context questions that inform the AI CTO and the report narrative without rolling into the score.

    • 5 Q
      Are we positioned to win?
      Strategy, ambition, competitive position, time horizons.
    • 7 Q
      Is it actually working?
      Outcomes, ROI, adoption depth, measurable productivity gains.
    • 13 Q
      What could go wrong?
      Risk, governance, data exposure, regulatory, vendor, brand.
    • 12 Q
      Can we keep up?
      People, capability, tooling provision, speed of execution.

    The score formula

    Each pillar score is calculated on a 0–100 scale that maps level 1 → 0, level 3 → 50, and level 5 → 100. The overall AI Maturity Score is the simple average of the four pillar scores.

    pillar_score = sum((level - 1) × weight) / sum(4 × weight) × 100
    overall_score = average(pillar_scores)

    "Is it actually working?" carries fewer scoring questions than the capability-heavy "Can we keep up?" or "What could go wrong?". That's deliberate — each outcomes question naturally carries more weight in its pillar score, reinforcing that actual outcomes matter more than process maturity.

    Skip logic and N/A handling

    Two upfront filter questions adjust which questions you see:

    • Do you develop software in-house? If no, the AI-in-software-development question is skipped.
    • Do you have a customer-facing product? If no, the AI-in-customer-experience question is skipped.

    Skipped questions are excluded from both the numerator and denominator, so they don't drag the score down. Same for "Not applicable" answers on department questions — the department is simply not scored.

    "Don't know" answers

    Every scoring question includes a "Don't know" option. This is deliberately distinct from "Not applicable":

    • Not applicable means the question doesn't structurally apply (e.g. "we don't have a Marketing function"). These are excluded from scoring entirely.
    • Don't know means the respondent can't vouch for the capability. These are scored as level 1 - a gap, the same as if the capability genuinely isn't present.

    We score "Don't know" as a gap because maturity assessments measure demonstrable capability - absence of evidence is treated the same way auditors treat unproven SOC 2 or ISO controls. Excluding them would create a loophole where respondents could dodge hard questions to inflate their score. If a "Don't know" is really a visibility gap rather than a capability gap, invite the right colleague to firm it up and re-take the assessment.

    Named maturity levels

    Your overall score maps to one of four named levels.

    0–34
    AI Aware
    AI is on the radar but unsupported.
    35–54
    AI Adopting
    Tools are in place but not consistently used or measured.
    55–74
    AI Embedded
    AI is part of how the organisation works day-to-day.
    75–100
    AI Leading
    AI is core to strategy, products, and operations — and measured.

    AI Risk Rating

    The AI Risk Rating is a secondary callout — not co-headline with the maturity score. It's the average rubric level (1–5) across these seven risk-flavoured questions:

    • D1Do you have AI-specific data protection policies in place?
    • D5How are customer and sensitive data handled when AI tools are involved?
    • D6Is intellectual property protection embedded in your AI tool usage?
    • RG1Do you have a formal AI Acceptable Use Policy?
    • RG3How do you assess and manage third-party AI vendor risk?
    • RG5Do you have incident response procedures for AI-related issues?
    • RG6How do you monitor AI outputs for accuracy and bias?

    The average rubric subscore maps to a band:

    ≤ 2.5High
    ≤ 4.0Medium
    > 4.0Low

    AI by department heatmap

    Six function-specific questions (Marketing, Sales, Finance, HR, Operations, Technology) score each department on the same 5-level rubric. Department scores are not rolled into the overall AI Maturity Score — they're a parallel diagnostic cut. Choose "Not applicable" for any function your organisation doesn't have, and that department is excluded from the heatmap entirely.

    Ready to take it?

    About 30 minutes. Free StackUp account at signup.

    Start the assessment