🛠️Deployment Guide (TrusstGPT on AWS)

This guide outlines the steps required to deploy TrusstGPT into your AWS account following subscription through the AWS Marketplace.

Overview of the Deployment Process

Overview of fulfilment process

Step 1: Prerequisites

• Preparation: Before starting the deployment, ensure you have administrative access to your AWS account and permission to create and manage AWS resources. Customers are advised to not use the AWS account root user for any deployment or operations.

• Request Service Quota Increases:

  • Running On-Demand G and VT instances: Needs to be increased by 64 More details and a link to access the request page can be found here.

  
• Please follow this process to request and Claude 3 Haiku and Cohere Embed English model access in Amazon Bedrock. See TrusstGPT Architecture for details. Note: It is possible to utilise Amazon SageMaker instead of Amazon Bedrock, but this will introduce additional infrastructure costs, since the Amazon SageMaker instance is always on, whereas you only pay for what you use with Amazon Bedrock.

• Troubleshooting: For any deployment issues, consult the troubleshooting section or log a case with Trusst AI support.

Step 2: Subscribe to TrusstGPT in AWS Marketplace

Navigate to AWS Marketplace and subscribe to TrusstGPT. (please wait for Trusst AI to share AWS Private Offering before subscribing if applicable). Specify your desired plan by inputting "1" unit, which corresponds to your organization's expected volume of contacts to process with TrusstGPT.

Step 3: Registration

Once subscribed via AWS Marketplace, you'll be redirected to a registration page (if you are not re-directed, please click “Set up your account”):

This will then redirect you to getting started with Trusst AI.

1. Enter the primary point of contact's details.

1. (Optional) Provide details of your Technical Point of Contact:

1. Provide your AWS account details for deployment configuration and seamless billing integration:

1. Select whether to use Trusst AI's default network configuration (Yes) or to provide custom network configuration (No) - Refer to Network Considerations for details.

1. (Optional - Recommended) Provide Corporate Identity Vendor for SSO integration:

1. Review Summary and Submit

Step 4: Setup

In a separate tab, make sure that you're logged into the AWS Console for the Deployment account specified in step 3.3 above, and then come back to the Trusst AI page and click "LAUNCH CLOUDFORMATION STACK" or copy the link into a new tab.

Before launching the stack, make sure you are logged into the AWS Management console with a user that has permissions to create IAM roles.

Step 5: Establish Trusst

Launch the AWS CloudFormation stack. The template will pre-populate the required fields.

Once you have reviewed the remaining configurations, acknowledge the capabilities of the AWS Cloudformation by ticking the radio button on the last page and clicking "Create Stack":

Once the stack is deployed, you will see “CREAT_COMPLETE” in the associated stack (CDKDeploy) in Cloudformation stacks in the deployment account.

Step 6: Deploy TrusstGPT

Trusst AI will trigger the deployment of into the AWS account and region from Step 3. The entire process takes approximately 45 minutes - 1 hour. You can monitor the progress of this via AWS CloudFormation in the deployment account.

Welcome Email: Upon successful deployment, the contact email provided will receive a welcome message including a link to Trusst AI’s Resource Center. This resource offers guidance on leveraging TrusstGPT for optimal benefits tailored to your use case.

Following these steps will ensure a seamless deployment of Trusst AI into your AWS account, enabling you to start leveraging the power of AI for processing customer contacts effectively. For further assistance or inquiries, please create a support case in the Trusst Support Portal.