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

# System Appropriateness Assessment (clause 3.2)

> The six sections of a System Appropriateness Assessment, how signing works, and what guardbook requires before it will let you sign.

A System Appropriateness Assessment asks one question about a tool: is it appropriate for what your firm actually uses it for?

It sits under clause 3.2, and guardbook holds one per tool. Open a tool from the Tool Register and the assessment section is near the bottom of its record, showing **Start assessment**, **Continue assessment** or **View assessment** depending on where you are.

## The six sections

| Section | What it asks |
| - | - |
| **Purpose justification** | Why this tool, for this use, at this firm. |
| **Capability evaluation** | What it can and cannot do. Its limits matter as much as its strengths. |
| **Alternatives considered** | What else you looked at, and why you did not choose it. Including doing the work without AI. For every option, this tool included, note what is known about its environmental impact and its impact on stakeholders. Clause 3.2 names both. |
| **Risk analysis** | The risk of erroneous output and of bias, and the consequences if either happens. Connect it to your Risk Register. |
| **Data handling** | What goes in, where it goes, what the supplier retains. Clause 3.1 is the test to apply here. |
| **Vendor due diligence** | Your summary position on the supplier. The detailed clause 4.1 record lives in [supplier due diligence](https://help.guardbook.ai/tools-register/supplier-due-diligence.md). |

## Draft and sign

The assessment saves as a draft at any point, so you can start it, gather what you need, and come back. **Save draft** does not validate anything.

**Signing does.** guardbook will not let you sign until:

* all six sections are at least 50 characters
* the named accountable surveyor is complete: full name, an RICS designation of AssocRICS, MRICS or FRICS, and a membership number of at least six characters with no spaces
* you type the accountable surveyor's full name, exactly, into the confirmation box

When you sign, guardbook downloads a PDF of the signed assessment. Save it. You cannot download it again from the tool's page. The assessment also appears in your evidence pack.

> [!NOTE]
>
> The 50-character minimum is a floor against blank sections, not a target. Six sections of 50 characters is not an assessment. It is six sentences.

Once signed, the record is read-only and shows the date and the signatory on the tool's page. It cannot be edited or signed again. Use **Resubmit** to archive the signed version and start a new draft pre-filled from it. guardbook keeps the archived version on record and logs the resubmission in your audit trail, but you cannot open it in the app, so keep the PDF you downloaded when you signed it. Until you sign the new draft, the tool does not count as having a signed assessment, including in your evidence pack.

## Who should be named

The accountable surveyor is the person who stands behind the judgement that this tool is appropriate. In a small firm that is usually the principal. In a larger one it should be whoever owns the decision, which is often the same person recorded as the tool's system owner.

> [!WARNING]
>
> guardbook checks the shape of a membership number, not whether it exists or belongs to the person named.
>
> Nothing stops one person completing an assessment and entering a colleague's details. If your firm cares about that, and it should, put the rule in your AI policy.

## When to reassess

The standard does not set an interval for this one. Sensible triggers are a change in what the tool is used for, a significant change to the tool itself, a new supplier position on data, or the next review date you set on the tool record.


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