Step 6: Verify the Connection

Use the AIAgnostic Test page to confirm each part of the connection independently before testing end to end. You don't need the Generate with MaxAI button to be enabled to run these tests.

In the App Launcher, find and select AIAgnostic Test.

The page runs three separate tests, one on each tab.

Note: If you make any changes to the Altify LLM Config or Altify In-App MaxAI Settings custom settings, refresh this page.

Test the External Endpoint

On the HTTP Test tab, you can run a plain connectivity check on your external endpoint.

  1. In Named Credential, enter the named credential from your Step 3: Configure the LLM Provider (for example, [CompanyName]_AIAgnostic_LLM).
  2. Select POST for the HTTP Method.
  3. In Endpoint Path, enter the endpoint path from your Step 3: Configure the LLM Provider (for example, /v1/chat/completions for OpenAI).
  4. In Request Body (JSON), enter a request body. The following example is for OpenAI:

    {"model":"gpt-4o","messages":[{"role":"user","content":"What is the capital of Ireland?"}]}

    The following example is for Perplexity:

    {"model":"openai/gpt-5.6-sol","preset":"fast","input":"What is the capital of Ireland?"}

  5. Click Test.

A successful test confirms only that Altify can reach the endpoint over HTTP with the configured authorization. It doesn't yet involve the LLM provider configuration or any prompt data.

If the test returns the following error, check that Allowed Namespaces for Callouts includes ALTF in your Step 1: Create the External Credential and Named Credential:

The callout couldn't access the endpoint. You might not have the required permissions, or the named credential "[Credential_name]" might not exist.

Test the AI Agnostic Config

On the LLM Config Test tab, you can run a data test that sends a prompt through the configured LLM provider and returns its response.

The fields are auto-populated from your HTTP test.

  1. In Request Body (JSON), enter a request body. The following example is for OpenAI:

    {"model":"gpt-4o","messages":[{"role":"user","content":"What is the capital of Ireland?"}],"response_format":{"type":"json_object"}}

  2. Click Test.

Success here confirms that the provider, model, and other settings in the Step 3: Configure the LLM Provider are correct, and that the round trip to the provider works.

Test the Generate with MaxAI Callout

On the Integration Test tab, you can exercise the MaxAI REST endpoint end to end.

To test the button for Account Details:

  1. Select Account as the Object Type.
  2. In Object Id, enter an Account Id.
  3. In Child Object Id (Question Id), enter the Id of an Altify account question that has a Step 5: Configure the Altify In-App MaxAI Settings.
  4. Click Test.

To test opportunity insight generation, enter an Opportunity Id as the Object Id. To test account insight generation, enter an Account Id.

Run all three tests before you enable the Generate with MaxAI button for users. A failure narrows down whether the issue is connectivity, provider configuration, or the MaxAI callout itself. When all three tests succeed, go to Step 7: Enable the Generate with MaxAI Button.