FAQs

A list of commonly asked questions and answers.

Q: What is the purpose of the Guardrails Toolkit page?

A: The Guardrails Toolkit page allows users to create, evaluate, and deploy policies.

Q: How can I create or update rules in the Create tab?

A: In the Create tab, select whether you want to create or update a Rule. You can manually input examples, use AI-assisted examples via interactive chatting, generate examples with AI, or upload a spreadsheet of examples.

Q: How do I update Rules?

A: To simplify the user experience, all Rule creation is currently handled by Preamble.

Q: What functionalities are available in the Policy Marketplace?

A: The Policy Market provides Rules and Policies shared by other organizations. Here, you can browse existing coverage if you need additional controls.

Q: How many rules are needed for a policy?

A: You can have 1 or more rules to a policy.

Q: When do rules and policies start versioning?

A: The user can edit a rule and policy until published. so that means no versions will be created; everything is version 1. When it's published to the public Marketplace, and you want to make changes, a new version is created. So, any future changes AFTER PUBLISHING create new versions.

Q: What support options are available if I face issues with my AI or platform?

A: For support or product feature suggestions, contact [email protected] or visit the Preamble website. For platform issues, submit an email with your name, organization, a brief description of the issue, and, if possible, screenshots.

Q: What is the configuration process and time period?

The configuration for API deployment involves making simple adjustments to your API code, for which we provide all the necessary code snippets. The duration needed to configure your policies primarily depends on the number of policies you intend to implement and the extent of testing you desire. Typically, implementing a single NLP policy can be completed in under 15 minutes. For the AI Assistant, no technical configuration is required by the user/admin.

Configuring policies for API deployment or the AI Assistant is straightforward and efficient — a single policy can be set up in less than a minute.

Q: How do you protect trade secrets and client confidentiality during configuration and installation? Preamble contractually agrees to maintain the confidentiality of our customers' proprietary information, trade secrets, and other sensitive data. This includes a commitment to not disclose or use customers' confidential information for any purpose outside of providing the contracted services. We also restrict access to customers' confidential information and data only to those employees and contractors who need it to perform their duties. These personnel are required to sign non-disclosure agreements to reinforce the confidentiality obligations. Preamble also implements robust security controls, such as encryption, access controls, and logging, to protect the client's data during configuration, installation, and ongoing operations. Finally, we ensure that customers' data is logically or physically segregated from other customers to prevent unauthorized access or commingling of information.

Q: What is your customer service schedule?

We offer multiple support channels, such as email (normal weekday business hours), phone (normal weekday business hours), and self-service options like FAQs and knowledge bases (24/7).

More comprehensive customer support coverage is available for higher tier customers. This includes faster response times and access to specialized support teams.

In addition to reactive support, Preamble has proactive customer outreach and engagement such as onboarding assistance, periodic check-ins, and proactive issue resolution.

Q: What is the Trial Subscription?

Preamble's Trial Subscription is a complimentary 2-week access to the feature-complete VanguardAI Platform, Marketplace, and AI Assistant.

Q: What is the Free Tier Subscription?

The free subscription gives customers access to our VanguardAI Platform, Marketplace, and AI Assistant with limited feature availability.