Internal Deployment

Deploying to azure in (35 Min)

What you will need to get started

You will need a few things to create an internal deployment with AI Risk.

1. Administrative access to create resources and Entra ID Applications within your azure environment

2. API Keys and Endpoints for your AI Models

3. (Optional) SMTP Endpoints, usernames and passwords to send emails if you want to send emails through your internal servers

Once you have signed an agreement for the enterprise version of AI Risk you will be provided with a zip file including three documents.

1. Your licence file (To activate your environment)

2. ARM Template to Deploy (To create infrastructure)

3. ApplicationCreate.ps1 (To create authentication application in Azure Entra)

Creating an Azure Entra ID Application (10 Min)

1. Go to https://shell.azure.com

2. Sign in with an administrator account or an account with access to create applications in Entra

3. Select powershell from the prompt

4. Upload ApplicationCreate.ps1 file using the upload and download file buttons in powershell

   
5. Run ./ApplicationCreate.ps1 in the console You will be prompted for a name less than 11 characters all lowercase make sure this is not generic as it creates resources with this name

   ./ApplicationCreate.ps1

6. At the end of the process there will be three values displayed (Save these values Used in the next step

   1. Project Name

   2. Client ID

   3. Secret

Running the ARM Template (25 Min)

The ARM template provided by AI Risk will allow you to deploy all the nessesary infrastructure to configure your environment internally within your azure environment

1. Go to https://portal.azure.com/#create/Microsoft.Template

2. Click "Build your own template in the editor"

3. Copy and paste the json text from MainTemplate.json found in the zip file provided and click SAVE

4. You will be prompted for the following values

   1. Resource Group (Create a new group using your naming conventions) All resources will be deployed to this group.

   2. Region

   3. Project Name (Use the project name that was output from the powershell script)

   4. Licence File URL (Will be pre populated from our script)

   5. SQL Server User Name (any valid username all lowercase)

   6. SQL Server Password (This should be a strong password but can be later updated to use integrated authentication)

   7. SMPT Endpoint (Pre populated with AI Risks enpoint but can be customized to use your own)

   8. SMPT Username ( (Pre populated with AI Risks username but can be customized to use your own)

   9. SMPT Password (Pre populated with AI Risks password but can be customized to use your own)

   10. Azure AD Instance (Should alwasy be https://login.microsoftonline.com/ but can be customized based on the application created)

   11. Azure AD Client ID (Use the Client ID that was output from the powershell script)

   12. Azure AD Secret (Use the secret that was output from the powershell script)

5. Click Review + Create The script values will be validated

6. Click Create

Loggin in for the first time

The first user to login will be allocated the Admin Role and allow them to configure and send new invitations and configure agents.

Browse to the Front end server which will have the default url of http://[Project Name]-app-service-frontend.azurewebsites.net

Using the hamburger menu in the upper left corner browse to the Admin Portal to configure your endpoints.

• Azure Open AI Service

• Open AI

• Mistral

Next create your first Agent and Invite new users

• Creating an Agent

• Invite new Users