Create supporting images with AI

Check availability first

AI image tools depend on your profile’s beta setting and the workspace media policy. If Create with AI is missing or disabled, check User Profile and ask the workspace administrator about media access.

Source-based creation and new illustrations use GPT Image 2.5 Sunburst. Annotation mode keeps the source image and adds editable marks. Review generated parts, labels, and positions against your sources before sharing; revise the request or use annotations when source details must remain unchanged.

Step generation and AI media have separate allowances. Review the remaining media allowance and reset date rather than assuming text generation includes unlimited images.

1. 1

   Open the three-dot menu in the app and choose User Profile.

2. 2

   Enable Beta features and choose Save Changes.

3. 3

   Open your project’s Media editor. If Create with AI remains unavailable, ask your workspace owner to check AI media access and the remaining allowance.

Prepare images in advance or create one later

You do not have to create images one step at a time. Before generating your work instructions, use the cog in Sources to enable Generate Media. StepGen then plans useful images across the generated steps and creates them in the background. Use this route when you want the first draft to arrive with supporting visuals.

For an individual improvement after reviewing the draft, open Media → Create with AI. That route gives you direct control over a particular step, references, and prompt.

Open the cog in Sources

1. 1

   Open your saved project and select the Sources tab.

2. 2

   Add the source material you want the instructions and visuals to use, then wait for processing to complete. PDF images or video frames give annotation and source-based creation something concrete to work with.

3. 3

   At the bottom of Sources, choose the cog beside Generate or Re-Generate. Its tooltip is AI Special Instructions. The AI Generation Settings window opens.

Enlarge

Open the cog beside Generate or Re-Generate in Sources to prepare the image settings before starting the instruction run.

Enable Generate Media before generating the steps

Existing media is retained when automatic visuals are added. The planner may propose an additional helpful image even for a step that already has media. Turn Generate Media off and save if you do not want automatic images in a later instruction run.

1. 1

   In AI Generation Settings, select Generate Media.

2. 2

   Open Additional options to choose Allowed media and enter any shared visual guidance.

3. 3

   Choose Save Settings to store the configuration for this project. Saving settings does not start image generation.

4. 4

   Back in Sources, choose Generate and complete the normal instruction-generation flow. If instructions already exist, Re-Generate starts an instruction regeneration flow; use the Media editor instead when you only want to add an image without regenerating your steps.

5. 5

   Let the steps finish, then follow media progress. Images can continue processing after the text is ready. Review each completed attachment before sharing.

Enlarge

Select Generate Media to request supporting images with the instruction run. Save Settings stores the choice; Generate starts the workflow.

Choose Allowed media: how should AI use your sources?

Under Generate Media → Additional options, Allowed media controls the kind of visual the automatic run may create. Your workspace may offer only some of these modes. Choose the least change needed to make the action clear.

Enlarge

Additional options combines the source-use mode with guidance that applies across AI media in this project. This example uses Create from source images.

Give the whole set of images a consistent style

Special instructions for media generation applies to AI media in this project and accepts up to 2,000 characters. Describe the visual conventions you want repeated: arrow color, readable labels, a close view of the target part, or a clean background.

For example: “Keep the source style. Use navy arrows to show the target part. Keep labels readable and do not invent measurements.” For new illustrations, you could ask: “Use a simple training illustration. Show only the objects needed for the action. Keep the background uncluttered.”

Keep approved process facts in your source material or Text Notes. Visual instructions shape the image; they should not be used to invent technical details that your sources do not establish.

Enlarge

Choose Create any supporting illustration when a new visual is appropriate. The selected mode and your shared instructions work together.

Understand the other options in the same window

The Sources cog does not expose an image-quality selector. Quality choices for an individual image are in the Media editor’s Create with AI options and depend on workspace policy.

Review the automatically created images

1. 1

   When the text draft is ready, check the generation progress and media activity for images still queued or processing. You can leave the editor and return later.

2. 2

   Open the steps and inspect the automatically attached visuals. Confirm each one explains the corresponding action.

3. 3

   Open Media to inspect AI images in the library and their attachment status. If an image was saved but could not be attached, select it and add it to the intended step.

4. 4

   For a missing or unsuitable visual, use Create with AI for that particular step, or attach an existing source image. Avoid regenerating the whole instruction set just to fix one image.

• Some steps may not need a visual. Automatic planning favors instructional value over decoration.
• Annotation and source-based modes need usable source images. If you only supplied text, add visual sources or choose the illustration mode when available.
• The per-run limit and monthly workspace allowance can leave some steps without a new image. Check media usage before requesting more.
• If a request fails, use its recovery message. Check activity before submitting again so that you do not duplicate work.

Create an image for a specific step

1. 1

   Open Media and choose Create with AI when available.

2. 2

   Check the step whose instructions will guide the image. Choose suitable reference images.

3. 3

   Describe the desired visual or edit. For example: “Highlight sample block B with a blue arrow. Keep all labels readable.”

4. 4

   Under additional options, choose the permitted source-use mode and quality. Submit the request once.

5. 5

   Follow its status in the media activity or library. Review the completed image, choose the destination step, and attach it.

Enlarge

Start in the Media editor and choose the step your image should explain. The Create with AI option appears when your account and workspace have access.

Review the visual, not just the status

Completed means an image was produced, not that its technical content is correct. Check part count, orientation, labels, and any measurements against your real source.

You can leave the editor while images generate and return later. If a request fails, read its recovery message before retrying. Reuse an existing good image when it already explains the action.