All Use Cases

See your AI spend, then the savings

Every AI number you have been shown was somebody's estimate

Spend comes first, attributed rather than totalled. Then value in tiers, with what somebody confirmed kept apart from what was estimated.

Powered By

The Challenge

  • Every AI figure you have been handed was an estimate wearing a confident face
  • The saving is always presented before the cost, which tells you what the number was built for
  • An hourly job gets credited with hourly savings, and nobody mentions it replaced a weekly chore
  • Hours saved get converted into money as though somebody had checked those hours were redeployed
  • You are asked to approve a renewal on a number nobody can reproduce

The Runwork Solution

  • Spend comes first and attributed, because a value figure presented before an unexplained cost is not a figure
  • Value is split by how hard it is: what somebody confirmed sits apart from what was estimated, and the two are never added together
  • Each rate records who set it and when, so an assumption has a name against it rather than being anonymous
  • A run is credited with the manual work it replaced, so an hourly job that replaced a weekly chore credits weekly rather than 168 times
  • Every report carries a section on failed runs, unprofiled resources and how much of the total is estimated only

What your team gets

Spend first, and attributed

What AI cost, broken down rather than totalled, before any value figure appears. A saving presented ahead of its cost is a number built to persuade rather than to be checked.

The cap that lowers the number

A job running hourly that replaced a weekly chore is credited weekly. Crediting it 168 times would produce a much better figure and a much shorter meeting.

An assumption with a name on it

Each rate carries who set it and when. That turns "we assumed two hours" into a question you can ask a specific person, which is the difference between an assumption and a guess.

Cost per run

Spend against execution counts, so the question stops being whether AI is expensive and becomes what a given piece of automated work costs to run.

What we refuse to turn into money

Hours spent rebuilding something are never dollarized, and work that never existed before is shown as capacity created rather than as cost saved. Both would be easy and neither would be true.

The difference from a value assessment

The assessments consultancies sell are projections, modelled before anything has run and sold on potential. This is the other way round: the work runs first and the report says what it did.

A Day in the Life

The renewal lands on your desk. Somebody wants to keep spending on AI and the case is a number you cannot reproduce. You have seen four of these and every one led with the saving.

You open the spend first. The report does too, which is the first sign it was not built to sell you something: cost broken down before any value figure appears.

You look for the soft part, and it is labelled. Confirmed sits apart from estimated and the two are never added. You take the confirmed figure, which is smaller than the one in the email you were sent.

You ask where a rate came from. The assumption carries who set it and when, so the question goes to a person rather than into the air. That is the moment this stops being a marketing artefact.

You look for what is missing. Failed runs, resources with no profile, and how much of the total is estimated only. It is a required section, so its absence is not an option and its presence is not a favour.

And the number is smaller than the one you were promised. That is the point. A figure that still holds after your second question is worth more than a bigger one that does not.

Frequently Asked Questions

Why does spend come before value?
Because a saving presented ahead of its cost tells you what the number was built for. Putting cost first is the credibility gate: you see what AI is costing, attributed rather than totalled, and only then what the work is claimed to have returned. It is a worse way to sell something and a better way to be checked.
How do I know the saving is not inflated?
Three things in the method lower it rather than raise it. Confirmed and estimated are never added together, so the headline is the checked part only. A run is credited with the manual work it replaced rather than with how often it ran. And time spent rebuilding something is never turned into money at all. Any of the three could be relaxed to produce a better figure, which is exactly why none of them is.
Who set the rates behind these numbers?
Somebody in your company, and the report says who and when. What a task used to cost is drafted by AI and then confirmed by a person who knows the work, and nothing it drafts counts until somebody confirms it. That human confirmation is why the confirmed column means anything, and it is also who you ask when a rate looks wrong.
How is this different from a value assessment?
A value assessment is a projection. It models what you could save, before anything has run, which is why the engagement that produces it is sold on potential rather than on results. This is the other way round: the work runs first and the report says what it did. One you buy in order to make a decision. The other you read afterwards to find out whether it was right.
What is the weakest number in the report?
The estimated column, and the report says so rather than making you find it. Estimated value has not been confirmed by anybody, which is why it is never added to the confirmed figure and never appears in a headline. The other soft spots are named in the same section: runs that failed, and resources nobody has profiled yet, which means their value is not counted at all.

Related Use Cases

Start keeping what your team figures out.

Start a 14-day free trial. No credit card required.