How much faster can you learn or explain something if every complex source you touch (papers, courses, decks – you name it) turned into a scannable infographic? Modern image AI makes it possible.
I view them as highly efficient “visual knowledge compressors”. Since Google’s “Nano Banana Pro” we got the missing link: solid text handling and visual reasoning. This means outputs are not “prettier noise” but logically sound. Try it for your own storytelling in low-threshold tools like Google’s Gemini Chat or NotebookLM and OpenAI’s latest ChatGPT Images. (When other providers finally get to that level, I’ll add them.)
But these simple tools are still limited in how much control you get over the outcome (content, look and feel). The problem: infographics have thousand faces. Timelines, dashboards, mind maps and more. If you do not set (or even know) the right parameters, the AI will improvise. Sometimes that is great for creativity, often it just misses your point. (Like always with GenAI: “Garbage in = garbage out”.)
Now, I’ll share my universal prompt template which you can adapt for any infographic type (with your tool of choice). I break the prompt down into its building blocks (logical chain of purpose, format, style, content etc.) with examples. Also, I’ll share some do’s and don’ts from my experience to give you a headstart.
Table of Contents
The Prompt’s Building Blocks
Think of this as a “menu”, not a form that you must fill end-to-end. No “one size fits all”. Just pick what’s already clear in your head, leave the rest open and iterate your way to your desired result in tandem with the AI:
1. Topic, purpose & audience/channel
Specify what the infographic is about, why you’re making it, who needs to “get it” and where it shows up. “Form follows function”. Examples: “AI roadmap for non-technical executives on LinkedIn”, “Quarterly KPI snapshot for senior management in a slide deck”.
2. Data / content basis & sources
What material or data should appear and where they come from. Plus, relevant context about the data (e.g. are these figures real or just indicative). Examples: Link to blog post, pasted XLS table, attached research PDFs, “use realistic mock data and flag it” etc.
3. Infographic framework / type
The overarching structural pattern that fits the story and data. (Consultants love this one.) Examples: Roadmap, funnel, flowchart, dashboard, tree map, mind map, juxtaposition (fancy?) etc. Endless possibilities…
4. Narrative flow & sections
How the viewer should (be) move(d) through the content from first glance to final takeaway. Examples: “Situation → Complication → Question → Resolution”, “Past → Present → Future” etc. – timeless classics.
5. Layout & aspect ratio
How big the canvas is, how it’s oriented and how “sharp” you like it. Examples: 16:9 wide-screen, 9:16 portraits (typical); 4K vs. 1080p (depends on “how much” there’s to show).
6. Text style & density
How much text you want and how it should appear. Examples: Shorter headings or one-liners only vs. longer explanations, business tone vs. edutaining, “minimal jargon”.
7. Visual style, colors & mood
Define the overall “look and feel” so the graphic appears intentional and “on brand” – not like a random template. Examples: Clean corporate, dark neon, soft pastels, isometric gradients (for techies), “color palette: [hex codes]”. Get creative.
8. Specific elements & constraints
“In scope vs. out of scope”: tell the AI what details (e.g. certain text phrases or visual cues) must be in the image vs. avoided. Examples: Chosen hero title, particular icons / shapes, reserved logo space, disclaimers, “no faces” etc.
9. References & examples for inspiration
If available, show the AI some concrete visuals as a clear steering example (aka “in-context learning”). Examples: Rough sketches, screenshot of a dashboard you like, slides in your intended style or brand guideline PDFs.
The Curtains Rise For… The Prompt!
Prompt Template (click to show/hide – and customize with [your inputs]):
“Create an infographic with a clear structure, accurate data relationships and crisp, readable text and no clutter. Use these inputs:
Topic, purpose & audience/channel: [what it is about, who it is for, where it will be used, e.g. executive slide, student syllabus, LinkedIn post].
Data / content basis & sources: [key facts, examples and links/files/notes; mark which numbers are exact or illustrative and which matter most]
Infographic framework or type: [e.g. timeline, funnel, flowchart, matrix, dashboard, tree map, bento, mind map, comparison or let the model choose a fitting structure].
Narrative flow & sections: [story path such as Situation → Complication → Question → Resolution, Past → Present → Future, plus the main sections].
Layout & aspect ratio: [aspect ratio like 16:9 (horizontal) or 9:16 (vertical); resolution such as 4K or 1080p].
Text style & density: [headers only vs shorter/longer explanations; tone such as professional, friendly, edutaining; amount of jargon].
Visual style, colors & mood: [e.g. clean corporate, dark neon HUD, minimalist, isometric; brand palette or hex codes].
Specific elements & constraints: [must haves like certain phrases, visual cues like call out boxes or logos; constraints such as no faces and no photo-realism].
References & inspiration examples: [optional sketches, exemplary infographics or dashboards; relevant style guidelines].
Plan the composition from these inputs so all elements are logically consistent and there is enough breathing room to prevent visual overwhelm. Create one image that follows this plan.”
For your inspo, I created this (quite dapper) exemplary infographic about the history of AI (with Nano Banana Pro). It’s based on my corresponding article. You can find the full prompt in the box below the image.
Click to expand for the example prompt
“Create an infographic with a clear structure, accurate data relationships and crisp, readable text and no clutter. Use these inputs: Topic, purpose, audience & channel: Visual summary of my article about the history of AI. Data / content basis & sources: You can find and read the article here. Infographic framework or type: Timeline. Narrative flow & sections: Chronological starting from the beginning on top and then progressing to the bottom. Layout & aspect ratio: 9:16 portrait. Text style & density: Each milestone gets its own label. Not too much text overall. Visual style, colors & mood: Dark neon HUD style and with isometric gradients appealing to techies. Specific elements & constraints: Use the chapter headings from my article for guidance which periods to cover and how. References & inspiration examples: Orientate yourself with infographic design best practices. Plan the composition from these inputs so all elements are logically consistent and there is enough breathing room to prevent visual overwhelm. Create one image that follows this plan.”
Do’s & Don’ts When Applying This
From my own experience and watching others, these small habits of smart human-AI-collaboration can make all the difference:
1. Do iterate in small steps.
Treat the initial output as a draft. Refine/edit your prompt; use local edits to fix specific parts instead of regenerating everything (or overthinking the prompt).
2. Do be specific on story, flexible on style.
Be clear about purpose, key message and data. If you are still exploring layout options or style, say so and let the model suggest a few styles.
3. Don’t fall for the polished first impression.
At ~80% quality we directly saw the “AI slop”. At 90%+ now it often feels “finished” and we stop questioning. That’s where risk sneaks in. Even my example above still contains “imperfections” requiring editing before serious use. So, let’s not get lazy.
4. Don’t outsource critical thinking.
Only feed the AI data you are comfortable sharing. Double-check anything critical (data, logic). Even the nicest graphic can be confidently wrong. You are the final filter.
Wrap-up: Tech Is The Canvas. You Are The Artist.
Equipped with this template, you can now use image generators to learn more efficiently or breathe life into your storytelling or data visualizations. I recommend you just try it out yourself for any complex topic you’re dealing with, so you can feel its power.
If you’re curious about the bigger picture of “visual information compressors”, check out this article. And if you are looking for more ways to put image AI to work (beyond infographics), I collected my favorite use cases here.
Your turn: I’d like to hear (and see) your personal “tops and flops”: what worked, what broke, where the AI surprised you in a good or bad way. Please share your experiences in the comments or reach out. If this prompt helps you, feel free to spread the word to other visual thinkers.
Cheers,
John
What do you think?