Creating AI-generated content for a mobile wellness experience
Timeline
3 months (Sep - Nov 2025)
Role
AI Content Designer & Developer
Skills
Midjourney, Suno, Airtable, Kling AI, Prompt Engineering
Project Context
What is U4RIA?
U4RIA is an AI-led lifestyle app featuring meditations, music, recipes, and affirmations, all generated using AI tools.
My role
During my internship, I created AI-generated content across multiple formats: social posts, recipe images, splash screens, affirmation audio, and short-form videos. I executed an end-to-end AI content pipeline, translating brand direction into structured prompts, managing iterations, and producing platform-ready assets. This required balancing creative intent with technical constraints and tight timelines while maintaining brand consistency across hundreds of generated outputs.
The AI Content Pipeline
PIPELINE
Process
Here is the systematic process I applied across all content types:
Define Intent & Constraints
Clarified the goal (educate, inspire, promote), emotional tone (calm, energizing), and format requirements before generation. This ensured outputs aligned with brand direction from the start.
Translate Intent into Structured Prompts
Used ChatGPT to convert creative direction into precise and descriptive prompts for Midjourney, Suno, and Kling AI. Iterated on prompt language to reduce ambiguity and improve output predictability.
Generate & Evaluate Outputs
Produced multiple variations and evaluated based on brand alignment, clarity, and technical quality. Selected strongest outputs that best supported the intended message.
Refine & Optimize for Platform
Brought assets into Canva to adjust composition, timing, typography, and formatting for different platforms (Instagram, TikTok, in-app).
Manage at Scale
Used Airtable to version control prompts, enabling reuse and consistency across content sets while maintaining flexibility for variation.
Projects & Outputs
SOCIAL MEDIA POSTS
Final Designs
I designed Instagram posts using Midjourney to generate the visuals, then brought them into Canva to refine typography, adjust spacing, and add overlays. The goal was always readability and making sure everything felt cohesive with U4RIA's brand.
RECIPE IMAGES
Final Generations
I generated photorealistic recipe images that matched the ingredients and cooking style described in each recipe. To keep things consistent at scale, I built prompt templates in Airtable that I could reuse and tweak across different dietary categories like bowls, comfort pastas, smoothies, you name it.
APP SPLASH SCREENS
Final Designs
I created animated splash screens using Kling AI. The trick was designing the start and end frames myself in Canva first. This way, I could control the composition and just let the AI handle the transition in between. The result felt calm and flowy, exactly what U4RIA's brand needed.
AFFIRMATION AUDIO
Final Generations
I produced affirmation tracks in Suno, where the voice and background music get generated together. I iterated on things like vocal tone, pacing, and the overall soundscape to match different moods, whether we needed something energizing, grounding, or calming.
SHORT-FORM VIDEOS
Final Designs
For TikTok and Reels, I assembled looping video sequences by generating scene imagery in Midjourney, converting it to video, then stitching the clips together in Canva. I synced everything to the audio track while making sure the loops were seamless and the mood stayed consistent throughout.
What These Projects Taught me
REFLECTIONS 💡
Lessons Learned
Specificity is a design constraint
AI performs best when prompts act as constraints rather than suggestions. Clearly defining intent, inputs, and desired outputs reduced unexpected results and shortened iteration cycles significantly.
Constrain inputs to speed up iteration
By locking down elements like start/end frames, voice characteristics, or visual tone upfront, I gained more predictable, repeatable outputs across generations and reduced trial-and-error.
AI is an execution layer, not a decision-maker
I learned to make creative and strategic decisions before involving AI. Treating AI as a production tool, not a source of direction, resulted in higher-quality, more intentional outputs and maintained creative control throughout.
Systemization enables scale
Building a repeatable pipeline and managing prompts in Airtable allowed me to produce hundreds of assets consistently. The investment in process design paid off through efficiency and quality at scale.
