Assist Help Guide: Image Generation

This guide explains how to create and use the Image Generation feature within Assist 2.0 to produce branded visual content, such as ad concepts, storyboards, and scamps.

The process is divided into two main parts:

  • Creating an Image Generation Template: Setting up the rules, style, and branding for your AI.
  • Using the Image Generation Template: Chatting with the AI to generate and refine your images.

There are now two ways to create an Image Generation template in Assist 2.0: building it manually in Workspace Templates, or asking the AI to scaffold one for you in chat. Both routes end up in the same template editor, so you can always fine-tune the details afterwards.

1. How to Create an Image Generation Template (Manual)

This template defines the AI's behaviour, style, and output format.

Navigate to Templates: From the main Assist 2.0 dashboard, select Workspace Templates from the left-hand navigation menu.

Create New Template: Click the + Create Template button in the top-right corner.

Fill in Template Details:

  • Template Name: Give your template a clear, descriptive name (e.g., "Ferrero Scamp").
  • What does this template do?: Describe its purpose (e.g., "Create me different scamps for Ferrero Rocher").
  • Template Icon: Click Change icon to select a relevant icon and colour for easy identification.

Select Output Type: In the "Output Type" section, select Images. This will reveal the "Image Generation Settings" section.

Configure AI Instructions:

  • Welcome Message: Enter the first message the AI agent will say to the user when this template is activated (e.g., "Hi, I am here to help you generate scamps for...").
  • AI Instructions: Provide a detailed prompt that defines the AI's persona, role, and task. This is the most important step for guiding the AI (e.g., "You are a Creative Concept & Design Agent specialised in luxury confectionery advertising... You must follow the approved layout structure...").

Configure Image Generation Settings (optional but recommended):

  • Template: Upload a base image (e.g., a PNG or JPG) that acts as a fixed layout. This helps the AI understand where to place permanent elements like logos or text blocks.
  • Mask: Upload a transparent PNG file (with the same dimensions as your template) to define where the AI is allowed to generate content. Transparent areas are where the AI can work; opaque areas are "masked off."
  • Training Examples: Upload up to 15 existing images that match the desired artistic style, branding, and quality. This trains the AI on the "look and feel" you want to achieve.

Set Image Output Settings:

  • Orientation: Choose from Auto, Portrait, Landscape, or Square.
  • File Format: Choose from Auto, PNG, JPG, or WebP.
  • Background: Choose from Auto, Transparent, or Opaque.

Set Visibility:

  • Public: Anyone in your organisation can see and use this template.
  • Private: Only you can use this template.

Save the Template: Click Update Template (or Save Template if creating from scratch) to finish.

2. How to Create an Image Generation Template via Chat

If you'd rather not build a template from scratch, you can ask the AI to create one for you conversationally.

  • Start a chat with Assist and describe the template you want — for example, "Create me an Image Generation template for Ferrero Rocher scamps, with a luxury confectionery style."
  • The AI will scaffold a new Image Generation template based on your description, including a suggested name, welcome message, and AI instructions.
  • Once created, the template will appear in your Workspace Templates list. Open it to refine the AI instructions, upload a base template, mask, or training examples, and adjust the output settings as described in section 1.

This is the fastest way to get a working template in place — then polish it in the editor.

3. How to Use an Image Generation Template

Once your template is saved, you can use it within any project to start generating images.

Open a Project: Navigate to the project where you want to create your images (e.g., "Easter Chocolates").

Start a New Output: Click the + New Output button in the top-right corner.

Choose Your Template: A pop-up window will appear. Scroll through the list and select the Image Generation template you just created (e.g., "Ferrero Scamp"). Click Create.

Interact with the AI:

  • You will be taken to the Project Assist chat interface. The AI will greet you with the Welcome Message you set up.
  • On the right, an Image Grid will appear, ready to display your results.

Give Your Prompt:

In the chat box at the bottom, type a prompt describing what you want to see.

Example prompt: "include an easter bunny and give me a greeting message"

Review Generated Images:

  • The AI will confirm your request and show a "Generating images..." status.
  • Your generated images will appear in the grid one by one.
  • The AI will also provide a text summary of what it has created and how it followed your instructions.

Preview and Refine:

  • Preview: Click on any image in the grid to open a larger preview. This modal also shows the detailed prompt the AI used to generate that specific image.
  • Iterate: Continue chatting with the AI to refine your images. Type a follow-up prompt to modify a specific image.

Example follow-up prompt: "take image #1 and add more samples of the product"

The AI will regenerate the specified image and update it in the grid.

Did this answer your question? Thanks for the feedback There was a problem submitting your feedback. Please try again later.

Still need help? Contact Us Contact Us