Higher education · Working concept · Evaluation and decision framework
AI Cost, Use, and Outcomes
Campuses need to know whether an AI use is worth continuing, not just whether people use it. Measure Cost, Use, and Outcomes against a baseline, then decide to continue, improve, expand, restrict, or retire. Designed for anything from one person’s task to a campus-wide platform, with effort scaled to the stakes.
The problem it solves
- Logins and licenses stand in for value
- Freed hours are counted as budget savings
- Tools renew by habit, with no baseline
- Each unit judges differently, so results can’t be compared
- Risk is checked once, then drifts
How it works
Full one-page diagram →
1 · StartA use caseWhat should AI improve, and compared with what?
2 · MeasureWhat happened?Cost, Use, and Outcomes.
3 · JudgeEvidence enough, risk acceptable?Quality, readiness, equity, risk, and evidence. Serious risks block use; other gaps set conditions.
4 · DecideWhat next?Continue, improve, expand, restrict, or retire.
Monitor and repeat. Decisions are revisited as results change.
Whoever uses it:IndividualTeamUnitInstitution
One evaluation, start to finish
Full case study →- 1 · Use caseA 12-person office drafts weekly meeting summaries with the campus AI assistant.
- 2 · Baseline35 minutes per summary, written by hand.
- 3 · CostNo new license cost. 30 min training each; 8 min review per summary.
- 4 · Use9 of 12 staff use it; 85% of drafts accepted after light edits.
- 5 · Outcome14 minutes per summary, including review. About 3.5 hours a week released.
- 6 · Evidence limitsFour weeks, small sample, no control group. No money saved.
- 7 · DecisionContinue. Released hours count as capacity, not dollars.
- 8 · Next reviewIn two terms, or sooner if the tool changes how it handles data.
Illustrative figures. You can start with one use like this; the Compass and Registry help at scale but aren’t required.
Choose your path
Leading the institutionThe answer, in two minutesThe main message, three points, and what it asks of leadership.Executive brief →
Running a unitIs AI worth it for my unit?Cost, Use, and Outcomes across the unit’s AI uses.Read the unit guide →
Using AI in my workIs AI worth it for my work?Cost, Use, and Outcomes for one person and one task.Read the guide →
Running an evaluationThe full frameworkEvery step, measure, and gate, in detail.Start section 01 →
Go deeper
The full framework
- 01Purpose: the problem and the answer
- 02Framework components
- 03From AI task to institutional value
- 04Measurement domains
- 05Capacity, savings, and value
- 06Multilevel decision matrix
- 07Evaluation lenses and decision gates
- 08Metric contract and aggregation rules
- 09Use-case scorecard and decision options
- 10How it fits with campus governance
- 11Putting it into practice
- 12The framework at a glance
Case studies
A working concept, open for improvement
It grew from seeing many claims about AI’s value in higher education and little empirical evidence behind them. It is a starting point for campuses to consider and adapt, not a finished standard. If you try it, share how you used it, what worked, and what should change.