Our approach
Assess. Benchmark. Advance.
A single progression from evidence to improvement. Each stage is designed so that what is measured can be acted upon, and what is acted upon can be measured again.
- Three stages
- Research principles
- Built with universities
Assess
Understand readiness and capability gaps.
Assessment is built around performance in realistic tasks rather than self-description. A student is asked to complete work with AI tools available, then to evaluate, correct and account for what those tools produced. This distinguishes capability from familiarity, and judgement from confidence.
- Task-based rather than attitudinal
- Domain-level results, not a single composite score
- Instrument behaviour examined alongside student performance
Benchmark
Build meaningful comparisons and institutional insights.
A number in isolation answers very little. Results are read against a comparative frame — across disciplines, cohorts and, where the evidence supports it, institutions — so that a pattern can be interpreted rather than merely reported.
- Within-institution comparison across faculties and programmes
- Cohort-to-cohort tracking as programmes change
- Sector comparison only where datasets are comparable and consent permits
Advance
Inform learning, development and measurable improvement.
The purpose of measurement is improvement. Findings are translated into specific implications for curriculum, assessment design, teaching practice and employability support — then measured again to see whether the change had the intended effect.
- Structured debrief with academic leadership
- Implications mapped to curriculum and assessment design
- Repeat measurement to evidence progress
Data flow
From a task to a decision.
Five stages, each writing evidence the next one reads. Select a stage to see what is collected and what it makes possible.
Realistic work, completed with AI tools available. No opinion surveys and no tool tutorials.
Diagram of the intended method. Stages and instruments are finalised with founding partners, subject to institutional ethics approval.
Evidence base
A benchmark is only as good as its evidence.
Before we measure anything, we read what has already been established. Every domain in the FutureMeasure index is drawn from published research on what employers, educators and standards bodies already agree matters — then examined for whether it can be observed in real work, rather than merely self-reported.
Method illustration · provisional domains
Select a column for the reported figure
Published evidence · our index is in development
| Attribute | Share of employers (%) |
|---|---|
| Problem-solving | 88.7 |
| Teamwork | 78.9 |
| Written communication | 72.7 |
| Verbal communication | 67.5 |
The chart reports published findings, not our own results. It shows the kind of evidence we build on: attributes employers already name, which our assessment then has to observe in performance rather than infer from a survey.
Research basis
Built on research that already exists.
We are not inventing a new theory of readiness. We are consolidating established evidence into a single comparative instrument, so a university can see where it stands against a frame that is documented and open to challenge.
- 01
World Economic Forum
The Future of Jobs Report 2025
2025
Analytical thinking remains the most sought-after core skill among employers, with seven in ten companies considering it essential.
Informs — Judgement under uncertainty - 02
National Association of Colleges and Employers
Job Outlook 2024
2024
88.7% of employers look for evidence of problem-solving on a graduate résumé, and 78.9% for the ability to work in a team.
Informs — Applied problem-solving - 03
UNESCO
AI Competency Framework for Students
2024
Positions students as AI co-creators and responsible citizens, with critical judgement of AI solutions at the centre of AI literacy.
Informs — Critical evaluation of AI output - 04
Stanford Institute for Human-Centered AI
AI Index Report 2025
2025
78% of organisations reported using AI in 2024, up from 55% the year before — adoption now precedes demonstrated capability.
Informs — Tool fluency in real work
FutureMeasure is independent of these organisations. References identify the published evidence base informing the method, not endorsement of it.
Research principles
The standards we hold the work to.
- 01
Methodology stated openly, so findings can be interrogated rather than accepted
- 02
Assessment design that reflects real work, not what is convenient to score
- 03
Consent, minimum necessary collection and careful stewardship of student data
- 04
Development with academic collaborators, including review before publication
- 05
Reliability and validity examined before any public claim is made
- 06
Interpretation that names its limits instead of hiding behind a ranking
What we will not do
Restraint is part of the method.
- —
Publishing validity claims before the evidence exists
- —
Reducing human capability to a single ranking for its own sake
- —
Selling certification before the instrument has been examined
- —
Treating tool familiarity as a proxy for professional judgement
Founding partnership
Interrogate the approach with us.
The most useful conversations are the ones where the method gets questioned. We welcome that.