What does Delegation Audit map?

It maps a real working week, prices the hours honestly, and helps decide which work moves down so the people you already have get more of their actual job back.

What happens to a task that cannot be described concretely?

It is not graded, not estimated, and does not reach the page.

Kit 10 · Workload mapping

Delegation Audit

Map a real working week, price the hours honestly, and decide which work moves down, so the people you already have get more of their actual job back.

Authority rule
The owner sets every verdict. AI asks the questions and does the arithmetic; a person decides what moves and what stays.

Field sheet

Find out where the week actually goes, before deciding anything moves.

Delegation Audit captures a week in the owner's own words, grades each task one question at a time, and produces a single printable page: what software can carry, what moves down to the person who owns that outcome, and what genuinely stays with the owner. Nothing on the page removes a person from the business.

Worked exampleIllustrative workflow

Current dragA fictional 14-person overhead-door company: the owner works a 52-hour week and cannot name where the time went. Names, hours, and figures are invented for demonstration.

Control addedEvery task is asked one question first: does this happen on a screen, or out in the world? Work that happens in the world is recorded as physical and can never receive an AI verdict.

System preparesFour fixed questions per task: concrete action, hours per week, what breaks if it is wrong, whose judgment it needs, then a lookup table assigns one of five verdicts and the arithmetic is shown line by line.

Human releaseThe owner confirms or overrides each verdict, states an effective hourly rate or declines to, and releases the one-page audit with reclaimed hours split into what software carries and what moves to a named person.

Processing boundary

Local by default. External by decision.

The operating record remains useful without a connected service. External assistance is limited to approved context and explicit tasks.

Runs locally

The captured task list in the owner's own words
The graded working file with hours and verdicts
The five-verdict table and the non-weekly conversion rules
The un-finished queue and its unblocking questions
The finished audit page and its branding folder

May run externally

Asking the interview questions one at a time
Organizing and restating captured tasks for confirmation
Drafting page copy from the graded working file

Never automatic: Grading a task the owner cannot describe, inventing hours, substituting a market rate for one the owner did not state, or claiming software performs work that happens in the world.

Ownership and hand-off

The workflow remains useful when the tool changes.

Client keeps the recordThe task list, the graded working file, and the finished page stay ordinary files in the client's own folder.
Every number is traceableEach headline figure is reproduced in an arithmetic block on the page, back to an answer the owner gave.
Verdicts stay overridableThe owner can overturn any verdict, and the override is printed on the page rather than quietly applied.
Tools remain replaceableThe audit is a plain HTML file that opens in any browser, prints on two pages, and emails as one attachment.

Human review checklist

Before the work moves.

Was every task asked whether it happens on a screen or out in the world?
Are the hours stated by the owner rather than estimated on their behalf?
Does the page carry zero dollar figures when no rate was given?
Do the reclaimed hours have a named destination: software, or a named person whose role grows?
Is every task the owner could not yet describe sitting in the queue instead of on a graded line?

Discuss this workflow Back to all kits