Chris Raroque
Chris RaroqueSep 12
Startups

App Branding Masterclass

16 min video5 key momentsWatch original
TL;DR

Hand-drawn original art combined with iterative AI refinement through ChatGPT creates uniquely personal app mascots that build emotional attachment far better than generic logos.

Key Insights

1

hand-drawn illustrations as foundationHand-drawn illustrations as a foundation for AI generation produces significantly more unique and stylistically consistent mascots than prompting AI directly without reference material.

2

break requests into individual promptsBreaking AI requests into individual, line-by-line prompts preserves the original style far better than asking ChatGPT to execute multiple modifications simultaneously.

3

emotional attachment over logosApp mascots create emotional attachment and connection that abstract logos cannot achieve, similar to why sports teams use mascots to build fan loyalty.

4

start new chats frequentlyWhen iterating with AI hits a wrong direction, starting a fresh chat is more effective than trying to correct course within the same conversation thread.

5

unlimited variations from base mascotOnce a strong base mascot is established, ChatGPT can generate unlimited variations for different app scenarios like empty states, loading screens, and user interactions.

6

names as changeable placeholdersApp names should be chosen quickly without perfectionism; they are changeable placeholders that can evolve as the product direction develops.

Deep Dive

The Mascot Advantage: Why Characters Beat Logos

Chris establishes that mascots create emotional attachment in ways that abstract logos cannot match. He uses the parallel of sports teams and their mascots to illustrate this principle. When users repeatedly see the same character—like Oliver the subscription monster sitting on receipts—they unconsciously build a personal connection to the app itself. This transforms productivity software from a cold piece of technology into something more relatable and memorable, making the user experience feel less sterile and more human.

The Three-Layer Creation Process

Chris's method combines human artistry with AI iteration in three distinct phases. First, he commissions hand-drawn original art (his fiancée drew their dog Luna as a base), which provides a unique stylistic foundation. Second, he feeds these illustrations into ChatGPT to generate variations and explore different directions. Third, once satisfied with a mascot, he creates specific scenarios using the base mascot—Oliver looking at receipts, holding a calculator, examining documents—for use as empty states and loading screens throughout the app. This layered approach ensures consistency while leveraging AI's efficiency.

The Critical Mistake: Asking AI Too Much at Once

Chris identifies a common error: giving ChatGPT multiple modifications in a single prompt (e.g., change the dog to a cat, swap the background, change the drink, alter the table color). While ChatGPT may technically follow instructions, it sacrifices the original style and reverts to generic AI-generated aesthetics. The solution is to break requests into individual line-by-line prompts, allowing the AI to maintain the base illustration's unique character while making incremental changes. This patience-based approach yields dramatically better results, particularly with ChatGPT-5, which has improved at style consistency compared to earlier versions.

Iteration Strategy: When to Persist and When to Reset

Chris emphasizes that mascot creation requires trial and error, sometimes taking 100 prompts to achieve the desired result. When a chat begins moving in an undesirable direction, he recommends starting a fresh conversation rather than attempting course correction, as image-based chats are difficult to redirect. During Oliver's creation, Chris iterated through multiple failed versions before completely reframing the concept as an abominable snowman-inspired character, which immediately yielded better results. This demonstrates the power of reconsidering the base prompt rather than stubbornly pursuing a failing direction.

App Naming Philosophy: Speed Over Perfection

Chris reveals his pragmatic naming approach: spend minimal time on names and treat them as changeable placeholders. His apps are named after his pets (Ellie the hamster, Luna the dog), and he spent only 2 hours brainstorming 'Subscription Monster' despite the critical role branding plays. He notes that his CRM app Mogul was originally called 'Karen' (after SpongeBob's computer) and Ellie was previously named 'Kronos' before changing them. The key insight is to move quickly, avoid perfectionism, and remain comfortable pivoting if the name or direction no longer fits the product evolution.

Takeaways

  • Commission hand-drawn original art as your AI foundation rather than starting with pure AI generation—it dramatically improves uniqueness and style consistency.
  • Break multi-part ChatGPT requests into individual prompts delivered sequentially; asking for too many changes simultaneously degrades style preservation.
  • When AI iteration heads in the wrong direction, start a completely new chat with a reconsidered base prompt rather than attempting in-thread corrections.
  • Use your finalized mascot to generate unlimited variations for app empty states, loading screens, and user interactions—this adds personality across the entire product experience.
  • Pick app names quickly without overthinking; treat them as changeable placeholders and remain flexible to pivot as your product direction evolves.

Key moments

0:45Why Mascots Matter More Than Logos

It's because it's way easier to get attached to a character than it is an abstract logo. When you add a mascot like I do to my apps, it becomes more than a piece of software.

3:00The Hand-Drawn Foundation

When you use original art as a foundation, it makes the mascot that you then generate in ChatGPT a lot more unique.

6:00The Critical Mistake: Too Many Instructions

If you ask it to do all of these things at once, the problem is it actually can't follow these instructions properly... it starts completely just not following the base style of the image you give it.

9:00Knowing When to Start Fresh

When it comes to image chats, my recommendation is to start a new chat frequently, especially when you notice it's going in the wrong direction because for image-based chats, it's usually pretty hard to correct course.

15:00Treat Names as Placeholders

I think just pick something, know that it could be a placeholder, and it can totally be changed in the future... I'm very, very comfortable changing it if needed.

Get AI-powered video digests

Follow your favorite creators and get concise summaries delivered to your dashboard. Save hours every week.

Start for free