> ## Documentation Index
> Fetch the complete documentation index at: https://docs.within.ai/llms.txt
> Use this file to discover all available pages before exploring further.

# Measure the value

> Recalculate against your baseline with comparable cases and the full cost of running the change, so the value claim holds up.

A change has delivered value when the people who own the outcome accept a measured result, not when it goes live. Measuring it means going back to the baseline you set in Discover and Define and recalculating with like-for-like evidence.

## Recalculate against the baseline

**1. Use comparable cases**

Compare the same kind of work before and after: the same segment, the same volume band, normal and peak cycles kept apart. A change measured on easier cases than the baseline overstates itself.

**2. Count the full cost**

Include what it takes to run the change: operating costs, the review burden on people, the exceptions that still go to a person, and the work that moved somewhere else rather than going away.

**3. Label every number**

Say where each figure came from and how confident it is. Combine the effects of several changes rather than adding them, because they overlap.

**4. State the limits**

Say what the measure doesn't cover and what would change it. A bounded claim that holds up is worth more than a large one that doesn't.

If the evidence isn't there yet, collect comparable proof first and recalculate before anyone makes the claim.

## Which numbers carry which case

The fields that carry a value case depend on what you're proving. A cost case lives on **Active Work Time × Annual Volume**. SLA, revenue and risk cases lean on Cycle Time and the narrative fields. See [How Advisor uses your process attributes](/user-docs/define/advisor-process-attributes) for which attributes carry each case and how precise they need to be.

## Who does what

| Role | Their part |
| - | - |
| **AI Builder & Deployer** | Supplies the cost, quality and version records |
| **Functional Champion** | Recalculates the value, and accounts for hidden work |
| **Executive Sponsor** | Accepts the bounded claim, and makes the next decision |
| **Process Performers** | Confirm the practical benefit, and the review burden |
| **Within** | Builds a traceable value review, and is clear about its limits |

When the value is verified, the first iteration is complete. What changes next is a decision, not a default. See [Maintain and improve agents](/user-docs/deliver/maintain-and-improve-agents).

## Where to go next

> **[How Advisor uses your process attributes](/user-docs/define/advisor-process-attributes)**
>
> Which attributes carry each kind of value case.

> **[Maintaining your process attributes](/user-docs/discover/process-attributes)**
>
> Raise confidence in the numbers a decision depends on.

> [!NOTE]
>
> **Need help?** Use the in-app chat — click the chat bubble in the bottom-right corner of Within (staffed 24×5) — or email [support@within.ai](mailto:support@within.ai).


This documentation is built and hosted on [Mintlify](https://mintlify.com), a developer documentation platform.