# Measuring released capacity

By Tommy Magee

Type: Framework
Canonical: https://tommymagee.com/resources/released-capacity

Time saved, minus time added, equals net capacity. Then check where that capacity went, and don’t count it as cash.

**Use this when:** A team says AI is saving time, and you need to know what that is worth and where it went.

## Why it exists

Time saved is the number everyone quotes and the one that holds up least. It leaves out the checking, the corrections, the upkeep, and the support an AI tool adds. It also treats a saved hour as if it were money. Lower spending is a saving. Released time is capacity, and capacity only counts if it goes somewhere.

## The framework

1. **Time saved: How much work did it take off?** Take the baseline time per unit of work and multiply by volume. Use the same definition of an acceptable result before and after.
2. **Time added: What did it add?** Count checking, corrections, upkeep, and support. Someone is doing these, even if no one is tracking them.
3. **Net capacity: What is actually released?** Time saved minus time added. That is the capacity. It is an estimate, so keep the range honest.
4. **Where it went: What happened to it?** Check whether it absorbed growth, avoided a hire, reduced spend, or simply evaporated. Report each separately.

## How to use it

1. Define a unit of work and an acceptable result, before the change.
2. Estimate the baseline time per unit and the volume over a period.
3. Estimate the time added by checking, corrections, upkeep, and support.
4. Subtract to get net capacity, and keep the range rather than one number.
5. Follow it forward. Record where the capacity went, and report it in those terms.

## Where released capacity can go

- It absorbed growth: the team took on more work without adding people.
- It avoided a hire: a role that would have been needed wasn’t.
- It reduced spend: something was actually cancelled or cut, which is a saving.
- It evaporated: the time freed up and nothing changed. This is common, and worth knowing.

## What it looks like

Picture a team that reviews a regular batch of documents. Before the change, each one took a person a known amount of time. After it, an agent prepares the review and a person checks it. The saving is the old time minus the checking time, and the corrections and upkeep come off that. What is left is the net capacity, and the team can then say whether it absorbed more documents, avoided a hire, or simply got a lighter week.

## Where it breaks down

Baselines are estimates, often made after the fact. Capacity that isn’t redirected disappears, so the last step matters most. And don’t report capacity as cash. Spending is only reduced when something is actually cancelled, and the two numbers belong in separate lines.

## Try it with AI

```
Help me measure the released capacity of one AI-supported workflow.

Ask me one question at a time. Define the unit of work and an acceptable result. Then ask for the baseline time per unit and the volume, and the time added by checking, corrections, upkeep, and support.

Calculate net capacity as a range, and ask me where the capacity actually went: absorbed growth, avoided a hire, reduced spend, or evaporated. Keep spending savings and released time in separate lines, and tell me which estimates I should check first.
```

Template: /templates/released-capacity-worksheet.md

Topics: AI ROI, Capacity, Measurement
