Capacity, savings, and value
In short: count hours as capacity, and dollars as savings only when spending falls.
- 1 · Task time reducedHours
- 2 · Capacity releasedHours
- 3 · Capacity redeployedHours, put to new work
- 4 · Financial benefit realizedDollars, only if spending falls
Report nonfinancial outcomes (learning, accessibility, service quality, equity) alongside any financial measures. Do not force them into speculative dollar values.
Acceptable evidence of redeployed capacity
From the use case inventoryTime saved is where many reports stop. To claim redeployment, show where the time went. Any one of these is enough if it covers a defined period.
Reallocated hours
Timesheets or schedules showing released hours moved to named work.
Added services
More appointments, courses, requests handled, or outputs, with stable staffing.
Backlog reduced
A queue that shrinks over a defined period with no added staff.
Work no longer deferred
Planned projects that started because capacity became available.
See Example 2, where released hours became 300 more counseling appointments a term.
Why hours rarely become dollars
AI reliably creates capacity. Spending falls only when someone decides to turn that capacity into lower costs. If the same people stay doing the same or more work, the hours are real but the budget doesn’t move. Recent industry surveys report the same gap: productivity gains with no change in spending.
Too small to harvest
Minutes saved across many people never add up to a whole role or contract.
Review eats the gain
Checking and fixing AI output take back part of the time saved.
New work fills the space
Freed hours absorb backlog or new requests, often unmeasured.
No one decides
Leaders never say where the time should go, so it disappears.
Pick a destination before launch
Before deploying, estimate the net hours released (after review time) and commit to where they go. Without this step, hours saved is a use measure, not a financial one.
- Option 1
Reduce spending
Retire a contract, cut overtime or contractors.
Counts as: Cash savings - Option 2
Avoid a cost
Hold a vacancy or absorb growth without a new hire.
Counts as: Cost avoidance - Option 3
Redeploy
Move the time to named work with an owner and a measure.
Counts as: Capacity, with an outcome - Option 4
Accept as capacity
Keep it as slack or service quality, and say so.
Counts as: Capacity, no cash claim
Capacity isn’t worthless. In mission-driven work it is often the goal: more students served, better service, people moved to higher-value work. Report it as capacity, not as savings.
Money terms, without double counting
Use these terms the same way across campus, over the same period, with assumptions stated. Never add them together when they count the same benefit twice.
| Term | Means | Counts as savings? |
|---|---|---|
| Total cost of ownership | Licenses, integration, infrastructure, training, review, governance, and upkeep | No, it is the cost |
| Gross benefit | Estimated value before costs | No |
| Net benefit | Quantified benefits minus attributable costs | Only the cash part |
| Cash savings | A verified drop in actual spending | Yes |
| Cost avoidance | Spending credibly avoided against a defined baseline | Yes, if documented |
| Released capacity | Staff time freed for other work | No |
| Opportunity value | What the redeployed time produced | No, report it separately |
Workforce effects to record
AI changes the work as well as the time it takes. Record these with Cost and the readiness check.
New review work
Hours spent checking and correcting AI output. Count it in Cost.
New skills
Training needed to use it well and to spot errors.
Changed roles
Who now reviews instead of producing, and whether that was agreed.
Who bears the burden
Whether the extra checking falls on particular staff or students.