Generating content

How to use the Knowledge Assistant to create documents, images, and other content.

Beyond answering questions, the Knowledge Assistant can also create content for you. Ask for a document, an image, audio, or video in plain language, and it returns a result you can refine and share.

What you can generate

The Knowledge Assistant can create several types of content for you.

  • Documents: Generate drafts, summaries, reports, and presentations in formats such as Word, PowerPoint, Excel, and PDF.
  • Images: Generate a picture that matches your description, without needing a separate design tool.
  • Audio (coming soon): Create voiceovers and other spoken audio from a script.
  • Video (coming soon): Create video from a description or script.

If you want to include specific formatting in your content, such as tables, charts, or dashboards, ask the Assistant to add them. If you create the same kind of content often, you can save your prompt as a custom command.

Some of this also comes as ready-made commands. Type /slides to turn a topic or a document into a PowerPoint deck, /image for an illustration or a photo-style image, or /infographic for a one-page visual with readable text, such as a job aid or a poster. To see every command, go to Assistant Commands.

When your content is ready, you can save it to your device to keep or share. Download documents straight from the chat, and right-click images to save them.

Generated content gives you a strong head start. Give it a quick review before you send or publish it, especially anything customer-facing or policy-related, so it says exactly what you want.

Generating images

Need a picture for an asset banner, a training slide, or a FlexDoc? Describe it to the Knowledge Assistant, which can create illustrations, realistic photos, and images with readable text, such as a poster with a title or an infographic with labels.

To generate an image, follow these steps.

Go to the Knowledge Assistant.

You can click the  Knowledge Assistant icon on any page, or Assistant in the main menu.

Start or continue a chat.

Describe the image you want.

The more detail you give, the closer the image will be to what you have in mind. It helps to mention the subject, the setting, and the style. For example, "Create a friendly illustration of a new employee's first day at the office, in soft, warm colors." You can also ask for a specific shape or text in the image.

Press Enter on your keyboard or click the  Send message icon.

When it's done generating, your image appears in the chat.

Getting the image you want

Use these tips to refine your request:

  • Shape and size: Ask for a square, landscape (wide), or portrait (tall) image, depending on where you'll use it. For example, landscape works well for an asset banner, and portrait for a poster. If you need a bigger image, ask for a larger size.
  • Detail: If the first version looks too simple, mention what to add, such as textures, lighting, or objects in the background.
  • Text: To include words in the image, such as a headline or labels on an infographic, tell the Assistant the exact text you want.
  • What to leave out: Specify what you don't want in the image, such as "Don't include any people." or "Leave out any text."
  • Refinements: After an image is generated, describe what to adjust in your next message, such as "Make the background blue." or "Add a coffee cup to the desk." The Assistant replies with an updated image in a new message, and the earlier one stays in the chat.

Want a chart or a photo that already exists instead? Here's how to ask:

  • Charts and diagrams: To turn your data or a process into a chart, flowchart, or diagram, ask for that by name. The Assistant builds it as a visualization you can interact with, rather than a picture.
  • Existing photos: Ask the Assistant to find images when you want photos that already exist, such as "Find photos of warehouse safety signs." It looks in your Knowledge Center first, then checks the web. To skip the Knowledge Center and search only the web, say so in your prompt, for example, "Search the web for photos of warehouse safety signs."

Viewing and saving an image

Click an image in the chat to expand it. Click the  Close icon to return to the chat.

To save an image to your device, right-click it in the chat and select Save image as (the exact menu option may vary slightly depending on your web browser).

When you expand an image from a web search, you'll also see an  Open original icon, which opens the full-size image in a new browser tab.

To copy the link to an image the Assistant generated or found on the web, right-click it in the chat and select Copy image address. Anyone with the link can then view the image in their browser for a limited time, even if they don't use Joy. After a short time, the link expires to keep your image private. That way, if the link gets forwarded or posted somewhere, no one can use it to view the image later.

Your images in Joy aren't affected by links expiring. Each time you open a chat, Joy refreshes the links automatically, so every image still loads.

Image Link works outside Joy for
Generated by the Assistant 30 minutes
Found in your Knowledge Center 30 minutes
Found in a web search 24 hours

Remember, the time limit applies only when someone opens the link outside Joy.

To share an image the Assistant generated or found on the web without it expiring, save it to your device and send the file instead.