How the grade works
What the value grade measures, the multiple behind the letter, what coverage means, and what the grade refuses to do.
What it measures
The value grade answers one question per use case: against what this use case costs a month, how much cash and how much time value did it return? Each use case in use gets a letter; the workspace gets one from the same arithmetic over every graded use case. The server strikes it; no screen computes a grade.
Three inputs, each with a name on the use case page:
- Allocated monthly cost: the use case's share of the tools and agents that power it. A shared cost divides across the use cases it powers.
- Cash saved, a month: money the owner says was saved. Negative stays negative.
- Time value, a month: hours freed times the workspace's value of an hour. Labelled not cash, because it is not.
Only a use case at Trying it or Regular use is graded. Planned, Paused and Retired are kept and shown, but a letter would claim a return that is not happening.
The multiple
Each bucket has its own multiple, and the page shows each one: $1.58 back per $1 of AI cost for cash, $0.69 of time per $1 of AI cost for time value. The letter comes from the buckets the grade reads (cash and time value by default) divided by the allocated cost, and that combined figure is never returned or shown: a reader who could see it would add it to the upside, which is the one sum the product exists to refuse.
The multiple becomes a letter by thresholds Oabo sets for the deployment. The thresholds are not published; the letter is a rank, and the sentence beside it carries the figures a reviewer can check.
Note
Scenario analysis shows a separate Cash ROI: (eligible cash value less allocated monthly cost) over allocated monthly cost, monthly, same scope. Time value and modeled upside are not in it. ROI, multiple and payback are different metrics, and the page says which one it is showing.
Coverage
The grade sentence says The grade counts 5 of 6 use cases: how many use cases in use hold a letter, out of how many are in use. A use case in use without a letter is listed under What's missing with the input it waits for. The sentence also carries the confidence of the whole page: Every benefit figure is Estimated until evidence is linked.
Coverage on the cost side is separate and is explained in Cost coverage: spend that no use case explains is shown, never silently allocated.
Not graded yet
| You see | Meaning |
|---|---|
| Not graded yet: add what it saves or earns | A value input is missing. The sentence names the first one. Unknown is not $0. |
| is not graded yet: add the AI tools you use and what they cost | Spend is unknown, so there is no denominator. The grade waits. |
| Graded once it is Trying it or Regular use | The use case is Planned, Paused or Retired. |
| No grade: possible extra revenue only | Its only benefit is modeled upside, which is beside the grade, not in it. |
| Counted in hours | Time was saved but has no dollar value, by the owner’s choice or because no value of an hour is set. The grade uses cash alone. |
| F. A result is negative. | The figures are in and the return is below cost. A negative keeps its minus sign; it is graded, not hidden. |
What it never does
- Add cash, time value and possible extra revenue together, anywhere, including the PDF.
- Turn hours into cash. A loaded hourly rate makes time value, which is labelled not cash.
- Count possible extra revenue. It is shown beside the letter with its own multiple and the words an estimate, not money received.
- Render an unknown as $0 or as F. Unknown, zero and negative are three states and look different.
- Blend in OKR attainment. It is its own figure, 0 to 1.0, beside the value grade and never added to it.
- Prove causality. A letter ranks a return the owner claims; evidence states say how far that claim has been checked.