Azure OpenAI
Requirements
Azure OpenAI API Key
Admin or better access to SurePath AI Admin interface
Create an Azure Connector
Login to the SurePath AI admin interface
Navigate to Connectors -> ADD CONNECTOR
Add a recognizable name for the connector
This will be used for selection when adding a GenAI model later
Select Azure OpenAI from the Type field.
Enter the API key from the vendor
Enter the Account Name
SAVE the form
You can now go to section Add and Enable your Private Model of this document to add and enable the private model.
Google Gemini
Requirements
Google Gemini API Key
Admin or better access to SurePath AI Admin interface
Create an Azure Connector
Login to the SurePath AI admin interface
Navigate to Connectors -> ADD CONNECTOR
Add a recognizable name for the connector
This will be used for selection when adding a GenAI model later
Select Google Gemini from the Type field.
Enter the API key from the vendor
Enter the Account Name
SAVE the form
You can now go to section Add and Enable your Private Model of this document to add and enable the private model.
AWS
NOTE: You MUST enable specific models in your AWS account before you can use them with the Portal. Follow the instructions below.
Requirements
Adequate permissions to enable Bedrock in AWS and create ARN access permissions
AWS account number
Admin or better access to SurePath AI Admin interface
Quick Look
There are three main functions you will need to perform to accomplish this HowTo.
Enable GenAI Model in your Cloud
Most cloud environments require you to approve the use of an LLM since this is additional licensing and will add cost.
Create a Connector
This enables SurePath AI to communication with your cloud environment to add and setup up permissions so we can communicate with your new LLMs once created.
Add and enable your Private Model
This is a simple matter of selecting the model you want to use, and changing the policy so that it's used.
Enable GenAI Models in your Cloud Environment
In this section you will enable Bedrock in your AWS environment. This will enable SurePath AI to communicate with new GenAI foundation models (FMs) that will be the basis for your Private Model at portal.surepath.ai.
Login to your AWS account
For best performance, select the region us-east-1
Access the Bedrock main page
Click on the Get started button
Read the information about required access and click Managed model access when ready.
Review AWS Bedrock Model Terms and other text on the screen. When ready, click the Enable all models button.
SurePath AI currently supports all of the ANTHROP\C models.
Note: This following may not be required and may not show on the screen depending on how your environment is configured.
Access the Overview Bedrock page at https://us-east-1.console.aws.amazon.com/bedrock/home?region=us-east-1#/overview
Enable the ANTHROP\C models by using the Request model access button and follow the prompts.
Create an AWS Connector
A SurePath AI Connector will be created to access the AWS environment where you just enabled the AWS Foundational models.
Login to the SurePath AI admin interface
Navigate to Connectors -> ADD CONNECTOR
Enter an appropriate Name (you will select this later) we'll use AWS-Prod in our examples.
In the Type field select AWS
In the AWS Account ID put your account number (12 digit number)
This can be found here: https://us-east-1.console.aws.amazon.com/billing/home?region=us-east-1#/account
The AWS Region should be: US East (N. Virginia)
The SurePath Role ARN (permissions for us to access your environment) will be auto-populated. This can be changed if required, please contact your F-CTO assistance if this is desired.
Select the Launch a Cloudformation Stack link at the bottom of the page.
This will enable SurePath AI to access the Bedrock FMs that you enabled in the first step.
This will open your AWS Console.
Please read and review all of the permissions that we are allowing. This will be READ only on your Bedrock environment.
Follow the instructions and allow the changes.
Please don't change the ARN in the process unless you have contacted us or really know what you are doing.
After completing the Connector it will take a few minutes to move from PENDING to ACTIVE on the screen.
Add and Enable your Private Model
Once your Connector turns green and Active on the Connector page you are ready to add a model and activate it.
Access the Private Models page at: https://admin.surepath.ai/config/private-models
Select the Add Private Model tile from the screen
Enter a recognizable Name. Please note that the official LLM name will already be visible on the tile that is created, so you don't need to put it into the name like you see above. We'll use Default Model
Select the Connector you created in the previous step AWS-Prod
Select one of the Supported Models from the dropdown list
Access the Default Policy page at https://admin.surepath.ai/config/default-policy
Select the Portal Private Model from the dropdown list that you just created: Default Model
Allow access to the portal by turning Portal Enabled green.
Scroll to the bottom of the screen and click SAVE CHANGES
Access the new Private model
Your new private model can be accessed at https://portal.surepath.ai
You can also gain access by trying to reach a GenAI application that you are not authorized for.
Upgrading to a new/different Private Model
Changing the model can be done by adding a new Private Model or Connector and Private Model.
Changing the Portal Private Model to the new one in the dropdown list.