Skip to main content
Back to Newsletter
Issue #007

Can AI keep your brand characters consistent? Leonardo AI says yes

Maintaining visual character consistency across AI-generated marketing content has been a persistent problem. Here is what Leonardo AI's Character Reference feature actually solves.

Christopher How
Christopher How
Ask Chris How
2 min read
TL;DR — Key Takeaways
  • Leonardo AI's Character Reference feature generates images that maintain a subject's likeness across different scenes and poses
  • Biggest win: brand consistency at scale without re-describing your character each time
  • Works best with illustrated characters rather than photorealistic brand imagery
  • Start with a basic subscription and test on a new project before committing

If your brand uses illustrated characters — a mascot, a recurring figure in social content, a spokesperson style across campaigns — you know how hard it has been to keep them consistent when using AI image tools.

Most AI image generators produce a slightly different face every time. It works for one-off images. It does not work for building a recognisable visual identity. Marketers have tried elaborate prompting strategies, style references, and ControlNet workarounds. Most deliver inconsistent results at scale.

Leonardo AI's Character Reference feature is a meaningful step forward. Here is what it does and whether it belongs in your stack.

What it actually does

Character Reference lets you upload a reference image of a person or character. The model uses that image as an anchor and generates new images that maintain the subject's likeness — different poses, settings, and expressions — without you having to re-describe the character in every prompt.

  • Upload a reference image and set the strength — how closely the output matches the reference
  • Describe the scene or action you want — the character's likeness is maintained across the generation
  • The Anime XL model currently gives the strongest results for illustrated characters specifically

Results are not perfect — secondary details will still vary. But the consistency improvement over standard prompting is significant enough to save real time for teams producing character-based content at scale.

Who this is actually useful for

  • Brands with an illustrated mascot or character used across campaigns
  • Marketing teams producing high volumes of social content with a recurring visual style
  • Businesses building course or email content using character-based illustrations

If your marketing is primarily photography-based, this feature will not do much for you. The value is specifically in illustrated or AI-generated characters used repeatedly across touchpoints.

The practical verdict

Leonardo AI runs on a subscription model. Paid tiers start around USD $10 per month, which is reasonable for the volume most marketing teams need.

Test before committing: take your brand character, run it through five different scene prompts using Character Reference, and evaluate whether the consistency meets your brand standards. If it does, build it into your workflow. If it does not, you have spent an hour and a small subscription fee to find out — which is a reasonable research cost.

Should marketers AI this? Yes — if you work with illustrated or AI-generated characters at scale. It will not replace a designer, but it will remove a lot of the manual consistency work that currently falls on them.

The Bottom Line
  • Character Reference solves the AI image consistency problem for illustrated brand characters
  • Best suited to teams producing character-based content at volume
  • Not relevant for photography-based marketing
  • Test with five prompts before subscribing — the results will tell you quickly whether it fits your needs