Acquia Web Governance uses Artificial Intelligence (AI) and Machine Learning (ML) to enhance your experience and speed up your workflows. This document provides information about the Web Governance AI and ML features and instructions on how you can manage them.
AI and ML uses in Web Governance ¶
The Web Governance AI and ML features are designed to streamline your tasks, save you time, and provide more accurate results.
Policy Assistant¶
Policy Assistant is an AI-driven, conversational interface (chatbot) that lets you create custom policies and run them in every scan of your website. It is based on the Google Gemini family of LLM models.
Access to the Policy Assistant is included with the Policies module.
How the Web Governance Policy Assistant is trained
The Policy Assistant is trained by Google and does not train itself on any Web Governance customer data or websites.
Consequences if Policy Assistant is turned off or experiences outages
There is no immediate risk or loss of data if Policy Assistant is switched off for your account or if it experiences an outage. You can manually create policies in the UI, and any policies previously created manually or with Policy Assistant continue to exist and run.
ML model for automatic scan releases¶
When the scan is done on your domain and there has been a large change in the number of pages from one scan to the next, the ML model may determine that Web Governance should not immediately release the scan results to you. This is because there may be cases in which your site is undergoing maintenance and it is not in your best interest to have those scan results available. The Web Governance ML model analyzes the scan results and makes a decision on whether or the scan results can safely be released automatically or if they require review by a human.
All customers benefit from this feature unless they explicitly request to have it turned off.
How the Web Governance ML model is trained
The ML model is trained by Web Governance on data from real-world scans to help make informed and better decisions in the future about whether to release a scan or not. No identifiable customer or website data are used in these analyses.
Consequences if the ML model is turned off or experiences outages
If the ML feature is turned off, there is no analysis of discrepancies in page drops or increases from scan to scan. The risk is that if, for example, your website is under maintenance and Web Governance only scans one page, the scan results are made available to you and you would lose all decisions made from the previous scan of the full website.
How to disable AI and ML features¶
Web Governance respects your preferences regarding AI and ML-driven functionalities. If you prefer not to use any Web Governance AI/ML features, you can disable them for your account or organization upon request.
To make a request to turn off AI and ML features:
- Contact the Web Governance support team
- Include your account details and which AI/ML features you would like to disable.
- The Web Governance team will confirm your request and send a notification when the changes take effect.