Price the use case.
Not the AI label.

Put the promised minutes, review work, data exposure, and monthly premium on the same balance sheet.

Built from the claim: AI is useful in specific cases; the rest can become a reason to charge more.

Make the assistant earn its keep.

A useful feature survives explicit assumptions. A vague promise disappears as soon as review time and price enter the same equation.

Decision

EARNS ITS KEEP

This specific use case clears its premium with room for uncertainty.

$120.88net value each month

6.0useful min / use

$155.88gross monthly value

3.35gross min / use to break even

4.45xvalue / monthly cost

(minutes saved - review) x uses x 4.33 / 60 x hourly value - premium - data allowance

A claim becomes useful when it can fail.

The audit keeps the skeptical parts visible, so a negative result is as useful as a positive one.

Specific use case, specific frequency.

"Has AI" is not an input. Name the task and count how often it actually happens. The calculation changes when the use case changes.

Saved time minus the work AI gives back.

Review, correction, and cleanup are not free. Useful minutes are the gross promise after that overhead is paid.

The verdict exposes its own threshold.

Break-even minutes show exactly what must be true for the premium to make sense. Export that boundary and test it against real use.

A useful assistant creates measurable surplus after review, price, and exposure.

Value after friction

Keep the proof. Lose the pitch.

The export records your assumptions, outputs, formula, and verdict boundary in a plain-text decision brief.

Review-adjusted value, recurring costs and two decision boundaries

Your browser does not support video.

Read the explanation

The default local assumption is eight saved minutes minus two review minutes, six useful minutes per use. Twelve weekly uses times a fixed four point three three weeks gives fifty-one point nine six monthly uses. Six times fifty-one point nine six divided by sixty times an assumed thirty dollars per hour gives one hundred fifty-five dollars eighty-eight monthly value. The thirty-dollar premium plus five-dollar exposure allowance totals thirty-five, leaving one hundred twenty dollars eighty-eight estimated net. Common two pixels per dollar bars measure three hundred eleven point seven six, seventy and two hundred forty-one point seven six. None of these assumptions are live prices, verified productivity or a quantified privacy loss. The wallpaper preset assumes half a weekly use, four saved minutes, two review minutes and twenty-five dollars per hour. Two useful minutes times two point one six five monthly uses divided by sixty times twenty-five gives about one dollar eighty monthly value. The eighteen-dollar premium plus two-dollar allowance costs twenty. Exact source arithmetic gives one point eight zero four one six seven value and eighteen point one nine five eight three three dollars shortfall. Common ten pixels per dollar bars show about eighteen point zero four value, two hundred cost and one hundred eighty-one point nine six shortfall. This is a supplied scenario, not evidence that any actual product is worth this amount. The source says skip when net value is negative, test when it is nonnegative but below twenty percent of monthly cost, and earns its keep when net reaches that threshold. For an assumed one-hundred-dollar cost, gross one hundred gives net zero and gross one hundred twenty gives net twenty. At three pixels per dollar, bars are three hundred and three hundred sixty with a sixty-pixel twenty-dollar margin. The break-even gross saved minutes adds review time to cost divided by monthly uses times hourly value per minute. With no monthly cost, the value multiple is labelled no cost; even zero useful value can hit the source earns threshold. Labels are transparent heuristics, not buying advice or validated risk estimates.