Guides Canva published apr 12, 2026
This guide teaches you how to turn raw research into an editable infographic using Perplexity, Gemini, and Canva Magic Layers. You will move from a data-heavy topic to a strong first-draft graphic you can clean up and ship instead of rebuilding manually.
You will build a repeatable workflow that turns a topic, memo, report, or dataset into an infographic you can actually edit in Canva. The final output is not a perfect AI image; it is a strong first draft you can clean up fast.
Tell the research tool what the final format is going to be. “Infographic concepts” is much better than “data summary.”
Start in your favorite AI research tool. The source workflow used Perplexity's deep research. Tell the tool that the end goal is an infographic instead of asking for a generic summary. Ask for outliers, comparisons, and narrative angles.
Do deep research on [your topic]. Look for outliers and narrative angles. Our end goal is four infographic concepts recapping the most interesting data trends.
In the source example, the topic was Daytona 500 data. The goal was not to collect every stat; it was to find the strongest storylines that could anchor a graphic.
Once the research is done, ask the tool to turn each concept into an image prompt.
Give me a prompt for each concept that I can paste into an AI image generator. Make the data easy to copy and paste.
This gives you a prompt set with the story and the data already packed in. Some prompts will be usable right away. Others will be too long or too cramped. Treat them as a starting point.
Paste one of the prompts into Gemini and generate a few infographic options.
The source workflow also tested Midjourney, but Gemini handled the data-heavy prompts better. Download the strongest result as a PNG. If you use another image tool, the same rule applies: you only need a clean first pass you can edit later.
Pro tip: Text-based infographics tend to work best. Once the image has lots of tiny charts, thin lines, or crowded labels, the next step gets harder.
Open a new Canva project, upload the PNG, select the image, then go to:
Canva will scan the image and separate it into editable parts. In the source test, each image took about 30 to 60 seconds, and Canva could process more than one at once.
Now fix the parts AI got wrong. This is where the workflow becomes useful.
The source workflow found three common cleanup jobs:
Text-heavy infographics came out closest to ready. Mixed text-and-image layouts usually needed a few replacements. Dense charts were the easiest way to break Magic Layers, so keep the concept simple.
Keep going
Members get all guides, courses, live sessions, and $1,000+ in partner perks.