Setting up AI features (On premise steps)

This advice is for on-premise (self-hosted) clients who wish to use the Claromentis AI features, as some set-up is required before these will work.

(In contrast, SaaS clients have these features readily available and can opt in to enable AI features anytime from Admin > System > AI)

To allow Claromentis to securely use Google Cloud’s Vertex AI models (like Gemini 2.5 and Imagen 3) on your behalf, you need to create a restricted Service Account and provide us with its credentials file.

This guide ensures you provide the absolute minimum permissions required (Vertex AI User), ensuring your broader Google Cloud infrastructure remains secure.

Step 1: Set up a Google Cloud Platform account

1. Head to Google Cloud and set up an account with associated billing details.

Step 2: Select or Create a GCP Project
 

1. Log in to the Google Cloud Console.

2. In the top navigation bar, click the "Project Dropdown"

3. Select an existing project you wish to use, or click "New Project" to create a fresh one dedicated to this integration.

4. Note: Ensure that your project has an active billing account linked to it, as Vertex AI requires billing to be enabled.
   
    

Step 2: Enable the Vertex AI API
 

Before a service account can call the AI models, the Vertex AI platform must be activated in your project.

1. Using the search bar at the top of the console, search for "Vertex AI API"

2. Click on the "Vertex AI API market" result (provided by Google Enterprise API).

3. Click the blue "Enable" button. (If it says "API Enabled," you are already good to go)

Step 3: Create the Service Account
 

1. Open the main navigation menu (the hamburger icon in the top left) and go to IAM & Admin > Service Accounts.

2. Click "+ Create Service Account" from the top menu bar.

3. Fill in the Service account details:

• Service account name: Enter a recognisable name (e.g., our-software-integration).

• Service account ID: This will auto-generate based on the name.

• Description: (Optional) e.g., Allows external app to access Gemini models via Vertex AI

4. Click "Create" and "Continue". Do not click "Done" yet.

Step 4: Assign the "Vertex AI User" Permission
 

1. Under the "Grant this service account" access to the project section, click the "Select a role" dropdown.

2. Search for and select Vertex AI User (or use the filter to find roles/aiplatform.user).

3. Click Continue, then click "Done".

4. Your new service account will now appear in the list on your screen.
    

Step 5: Generate and Download the JSON Key File
 

1. In the Service Accounts list, look for the account you just created and click on its Email address.

2. Navigate to the Keys tab at the top of the page.

3. Click the "Add Key" dropdown menu and choose "Create new key".

4. Select "JSON" as the key type, and click "Create".

5. A .json file will automatically download to your computer.
    

⚠️ Important Security Warning: This JSON file contains the private key that allows access to your Vertex AI models. Treat it like a password. Do not share it via email, commit it to public repositories, or modify its contents.

Step 6: Upload the Key to Claromentis
 

1. Open your Intranet (your profile needs to be a sysadmin)

2. Navigate to Admin > System > AI Settings

3. Locate the Google Cloud Vertex AI JSON Key upload field.

4. Upload the downloaded .json file and save your settings.

5. Claromentis will run a quick verification test to confirm it can successfully reach the Gemini-2.5-flash and imagen-3.0 models, and your setup will be complete!

Step 7: Test the AI features

1. The features will now be available for use on your site.
    
2. Sysadmins of your site can toggle these on/off from Admin > System > AI Settings at any time
    
3. Test out the generative AI features as described here
    
4. Test out the AI search as described here (v11 - currently unreleased)