fix: clarify model configuration prompt wording

[v1212]

Summary

Clarifies the model configuration prompt options to reduce user confusion.

Relates to #8155

Before

? How would you like to configure model(s) for your agent?
  > Deploy new model(s) from the catalog
    Use existing model deployment(s) from a Foundry project

After

? How would you like to configure model(s) for your agent?
  > Deploy new model(s) (with new Foundry project)
    Use an existing Foundry project

Problems with the previous wording

1. "Deploy new model(s) from the catalog" — did not indicate that this path also creates a new Foundry project, misleading users who already have a project and just want to deploy a new model into it.
2. "Use existing model deployment(s) from a Foundry project" — implied only existing deployments could be used, but this path also allows creating new model deployments within the existing project.

The new wording accurately reflects that the choice is fundamentally about which Foundry project to use (new vs existing), not solely about the model deployment itself.

Related

• Bicep fix for creating model deployments on existing projects: fix: create model deployments on existing Foundry projects during provision Azure-Samples/azd-ai-starter-basic#67

[Copilot]

Pull request overview

Clarifies the interactive model-configuration prompt wording in the azure.ai.agents extension to better reflect that the decision is about which Foundry project to use (new vs existing), reducing user confusion (relates to #8155).

Changes:

• Updates the model-configuration choice labels in the main init flow to be Foundry-project focused.
• Aligns the init-from-code flow’s model-configuration choice labels to the same wording.

Reviewed changes

Copilot reviewed 2 out of 2 changed files in this pull request and generated no comments.

File	Description
cli/azd/extensions/azure.ai.agents/internal/cmd/init.go	Updates init prompt option labels to distinguish new Foundry project vs existing Foundry project paths.
cli/azd/extensions/azure.ai.agents/internal/cmd/init_from_code.go	Updates init-from-code prompt option labels to match the clarified Foundry project wording.

Comments suppressed due to low confidence (1)

cli/azd/extensions/azure.ai.agents/internal/cmd/init_from_code.go:505

• In this init-from-code flow, the prompt message/comment still frame the decision as configuring “a model”, but the options were updated to be about choosing a Foundry project and (potentially) multiple models (“model(s)”). This mismatch can confuse users and undermines the clarification this PR is aiming for. Update the prompt message/comment to match the new option wording (project-focused and consistent with init.go).

	// Ask user how they want to configure a model
	modelConfigChoices := []*azdext.SelectChoice{
		{Label: "Deploy new model(s) (with new Foundry project)", Value: "new"},
		{Label: "Use an existing Foundry project", Value: "existing"},
		{Label: "Skip model configuration", Value: "skip"},
	}

	var modelConfigChoice string
	if a.flags.projectResourceId == "" {
		defaultIndex := int32(0)
		modelConfigResp, err := a.azdClient.Prompt().Select(ctx, &azdext.SelectRequest{
			Options: &azdext.SelectOptions{
				Message:       "How would you like to configure a model for your agent?",
				Choices:       modelConfigChoices,

[trangevi]

Fine with me, unless we want to complete pivot to "Create a new foundry project or use an existing one", and then "create a new model deployment or use an existing one" if users select to use an exisiting project. @therealjohn

[therealjohn]

I think the right fix here is to decouple projects from models.

We also don't need to ask for a model unless its a resource defined in the agent manifest.

@v1212 can you pivot this to focus on that?

[v1212]

@trangevi @trangevi yes, I intended to decouple completely project selection/creating from model selection/creating in parallel, this PR is quick one to make existing flow easy to follow and avoid confuse.

I will work on a follow up PR to refactor in azd init, key changes will be like below, please suggest if any thoughts. feel free to leave this pr here as it can only help clarify before the refactor gets released.

? Select a Foundry project:

Use an existing Foundry project
Create a new Foundry project
--- If "Use an existing Foundry project" is selected ---
? How would you like to configure model(s) for your agent?
Use an existing model deployment
Deploy a new model from the catalog
Skip for now
--- If "Create a new Foundry project" is selected ---
? How would you like to configure model(s) for your agent?
Deploy a new model from the catalog
Skip for now

in total, 5 cases will be supported
if using existing project, then create/choose/skip model
if create new project, then create/skip model