All coursesAI guardrails for teams
Using the SAL as a guardrail

Using the SAL as a guardrail

Use the Site Activity Log to review AI-assisted changes before you publish.

Using the SAL as a guardrail

Use the Site Activity Log to review AI-assisted changes before you publish.

When AI is part of your workflow, the Site Activity Log (SAL) gives you a clear record of what happened on your site: what changed, who made the change, how it was made, and when.

This lesson is about making that review part of your workflow, especially after an AI-assisted session.

Access to the Site Activity Log requires a Platform plan. Visit Webflow Plans & pricing to learn more.

What to look for in the SAL

The SAL helps you answer the questions that matter before and after publishing:

  • What changed? Review updates to pages, styles, CMS items, components, settings, and more.
  • Who made the change? See which person was associated with the activity.
  • How was it made? Use AI attribution to understand whether the change came from manual work, a Webflow AI tool, or an MCP-connected agent.
  • Where should you look closer? Pay extra attention to broad or shared changes, like updates to styles, variables, components, or page settings.
A single SAL entry expanded to show affected elements and pages.

The third question, how was it made?, is what makes the SAL especially useful as an AI guardrail.

AI attribution helps you separate manual changes from AI-assisted ones, so you can review the scope of an AI session more clearly. It can also help teams understand how AI is being used across the site: who is using it, where it's showing up, and what kinds of changes it's supporting.

The Webflow Site Activity Log showing a list of recent changes, with AI attribution tags visible alongside human-attributed entries.

Use that context before you publish to check whether the scope of the work matches what you expected. Use it after publishing if something looks wrong and you need to trace what changed.

Querying the SAL with MCP

If your team is working with MCP, you can also query the SAL in plain language. Instead of scrolling through entries one by one, you can ask for a summary of what changed.

For example:

Give me a structured summary of everything that happened on this site during the last session. Group the changes by type, and flag anything that should be reviewed before publishing.

Depending on what happened in the session, the response can help you spot patterns faster: pages that changed, styles that were updated, CMS items that were added, or anything that may need a closer look.

You can keep a prompt like this handy as part of your team's pre-publish process.

A Claude conversation showing a structured summary response, with activity by contributor and change types grouped.

The staging connection

In an Enterprise publishing workflow, staging gives you one more checkpoint before changes go live.

Before publishing, you can review recent changes and click into any entry to open the related SAL record for more detail. Once the changes look right, you can publish to staging and even choose to publish a single page.

An Enterprise staging publish view showing recent changes listed before publishing, and the option to publish a single page.

A few habits worth building

The SAL works best when it's part of your routine.

  • Review the SAL before publishing after an AI-assisted session.
  • Filter by change type when you need to focus your review.
  • Pay extra attention to variables and styles, since those can affect many elements and pages.
  • Use an MCP summary when the session included a lot of AI activity.
  • Click into any staging change that looks broad, unexpected, or high-impact.

The point isn't to review every small change forever. It's to know what changed before it reaches your live site.

Ready to continue?

Click Complete & continue to wrap up with best practices for using AI guardrails across your workflow.

1

Intro

Coming soon

1

Background & preview
2:00
Background & preview
Coming soon

2

Managing AI access

Coming soon

2

Setting AI access for your team
3:29
Setting AI access for your team
Coming soon

2

AI within role limits
2:30
AI within role limits
Coming soon

3

Understanding AI changes

Coming soon

3

How far do AI changes reach?
2:30
How far do AI changes reach?
Coming soon

4

Tracking and reviewing changes

Coming soon

4

Site Activity Log
3:43
Site Activity Log
Coming soon

4

Using the SAL as a guardrail
2:00
Using the SAL as a guardrail
Coming soon

5

Wrap up

Coming soon

5

Best practices & resources
2:00
Best practices & resources
Coming soon

Course progress

0%

Assessment

Up next

Best practices & resources

Practical habits for safer AI-assisted workflows in Webflow.
Complete & continue
Complete course