Transforming Our Brainies: AI Image Manipulation Experiment - 5 / 5

Transforming Our Brainies: AI Image Manipulation Experiment - 5 / 5

Part 5: Find the best option - Asking Manus

In Part 2, Part 3, and Part 4 we tested twelve tools to enhance one of our Brainies, the text model. We decided in Part 1 to take inspiration from Agatha Christie to give more personality to the mascot.

The origin of this journey comes from Part 1, when we used Manus, an AI Agent, to analyze our mascots, based on a white paper from Duhno. Now, we are going back to Manus, with seven candidates to replace our current mascot.

Our approach

We asked Manus to use again the white paper on mascots to score every mascot, and find the top 3. Out of this top 3, we want recommendation to further improve them.

Here is our prompt: We have designed 7 potential replacements for our current mascot based on the analysis performed using a white paper on the ideal mascot. In the attached picture, the first one is our current mascot (top left corner), all the others are candidates. Based on the attached white paper, score every mascot on the 5 key traits of a successful mascot. Give us a ranking. For the top two mascots, give us instructions to make it even better, looking for a perfect score.

Results

Here is Manus’ scoring (top 3 and original):

Mascot	Recogni-zable	Consistent	Approa-chable	Relevant	Memorable	Overall
	5: Highly recognizable. Brain shape is clear. Very expressive.	4: Cartoonish style. Well-defined features.	5: Wide, genuine smile. Engaging eyes.	5: Directly relevant as a brain. Symbolizes intelligence.	5: Combines a brain with expressive, friendly face.	24/25
	4: Brain shape clear. Distinct, friendly facial features.	4: Cartoonish style. Defined facial features.	5: Large, friendly eyes, smile, and overall cheerful demeanor.	5: Directly relevant as a brain. Symbolizes intelligence.	4: The friendly face and expressive eyes.	22/25
	3: Bust shot. Focuses on expression.	4: Relatively easy to reproduce consistently.	4: Gentle smile and Thoughtful pose.	4: Deep thinking, mentorship, accumulated wisdom.	3: Memorability depends alignment with brand.	18/25
	3: Simplicity borders on generic, and it might not stand out.	4: The personality is not strongly defined.	2: Lack of facial features makes it difficult to convey emotion.	4: Relevant to themes of intelligence, thinking, or knowledge.	2: Generic appearance and lack of distinct personality or features.	15/25

Recommendation on top 3

Manus also provided recommendations to improve each of our top two:

Mascot	Recommendations
	Consistent: To ensure perfect consistency, especially with the detailed brain texture, develop a comprehensive style guide. This guide should include precise color codes, rules for how the texture is rendered at different sizes and in different media (e.g., print, web, animation), and clear guidelines for depicting its range of expressions and poses.
	Recognizable: While the brain shape and friendly face are good, consider adding a small, unique visual element that is always present, beyond just the book. This could be a tiny, distinct mark on the brain, a unique shape to its eyebrows, or a specific, subtle color pattern within its irises. Consistent: Develop a detailed style guide. Since its features are slightly simpler than previous one, focus on defining the exact proportions of its eyes, mouth, and head shape. Memorable: Enhancing recognizability with a unique visual element (as suggested above) will also boost memorability. Additionally, consider developing a signature pose or a very short, catchy animation/sound associated with it.

Style Guide: Agatha Brainy

Since Manus’ suggestion is a style guide. We asked for one: Give us a style guide for the chosen Mascot in two hundred words.

Here is the result with a few minor adjustments.

Introduction

Objective: To ensure consistent and appealing visual representation of Brainy across all applications. Core Visuals: Agatha Brainy is a cartoon brain character defined by its highly expressive, friendly face and detailed brain texture. Key features include large, bright blue eyes (with pupils and highlights for expressiveness), a wide, genuine smile often showing dimples, and clearly defined, symmetrical brain convolutions that add texture without being overly complex.

Color Palette

• Brain: Warm, light grey or a soft, pale pinkish-beige (e.g., #E0D8D5 or #F5EDE9). Avoid stark white or overly clinical greys.
• Eyes: Bright, engaging blue (e.g., #4A90E2) for irises, with white sclera and black pupils.
• Mouth: Pink/red interior (e.g., #E57373).
• Limbs/Accessories: Neutral dark grey or black for limbs (e.g., #4A4A4A). Book cover can be a friendly accent color (e.g., warm yellow #F8D96C or teal #4DB6AC); pages are off-white. Quill is dark grey/black with a lighter feather tip.

Shapes & Texture

• Form & Proportions: Maintain a rounded, full brain shape. Eyes should be large and prominent, occupying a significant portion of the face. Limbs are simple, stick-like, and secondary to the expressive head. Ensure proportions remain consistent, especially the head-to-eye ratio.
• Texture & Detail: The brain convolutions should be visible and add character but simplified for clarity, especially at smaller sizes. Avoid overly realistic or grotesque detailing. The texture should feel organic and approachable.
• Expressions: While the default is cheerful and intelligent, Brainy can display a range of positive emotions (curiosity, excitement, thoughtfulness) primarily through eye shape, eyebrows (if added stylistically), and mouth variations. Maintain a friendly essence always.
• Application: Strive for clean lines and clear forms. Ensure legibility at various sizes. A simplified version may be needed for very small applications, focusing on the core facial features.

Final Conclusion

In this series of articles, we can draw several conclusions:

Image Manipulation

There are many tools on the market that pretend to perform AI-powered manipulation, but only a few are easy to manipulate and produce good results. Each tool works completely differently and requires practice to refine a specific use case.

Branding Analysis

Manus, as AI Agent, can combine the power of a vision model, text analysis, and deep thinking. That gave us very good results in initial analysis and final scoring of the results. That was only possible because we had access to very detailed and insightful reporting on mascots. This framework was critical to the success of our strategic analysis. Finally, Manus was able to give us recommendation and execute on that recommendation by generating a style guide.

Next Steps

Now that we have a direction and a set of tools, we need to go through all eight of our mascots. Our target is to get at least 23 on the scoring system for each mascot in its domain. We also need to add all our brand elements in the evaluation. That was an oversight in the last part of this experiment.

We are Here to Empower

At System in Motion, we are on a mission to empower as many knowledge workers as possible. To start or continue your GenAI journey.