The future of AI content is not better prompts. It's better systems.
A lot of AI-generated content looks like AI-generated content. Not because the models are weak, but because the process around them tends to be light. People type a sentence into an image generator, get something back, tweak a few words, and try again. There's rarely a creative framework behind it, and no visual logic tying one output to the next. The result looks impressive in isolation and falls apart the moment you put two pieces next to each other.
I've been working with AI tools in creative production for a while now, and I keep seeing the same pattern. The bottleneck isn't the model. It's the gap between a clear creative vision and prompts that actually deliver it, consistently, across tools and scenes.
The quality problem
There's a lot of AI content out there right now, and much of it sits somewhere in the middle. Not because the technology can't do better, but because speed keeps winning over craft. Generating twenty images in ten minutes is easy. Generating twenty images that feel like they belong to the same project is hard.
This is the same problem that has always existed in creative production. Consistency requires a system, and in traditional work that system is called art direction. Someone defines the visual language, the lighting approach, the colour logic, and the texture palette. Every individual piece then gets produced within that framework. That's what makes a campaign feel like a campaign instead of a mood board dump.
AI tools don't have this layer by default because they're stateless. Every generation starts from zero. If you want consistency, you carry it across every prompt yourself. Detailed, structured prompts, written over and over, adjusted for each tool's syntax and quirks. It works, but it's slow, repetitive, and easy to get wrong.
What I built
I called it the Prompt Enhancement Engine (the project page covers how it fits into the rest of the work). You give it a creative brief and a few reference images. Before writing a single prompt, it generates an art direction layer: a structured interpretation of your brief that defines the lighting language, material logic, colour palette, and overall mood. The framework an experienced art director would set up before any production begins.
From that framework it generates a full set of prompts for image generation, image editing, and video. All of them come from the same visual logic, so when you change the brief, everything updates together.
The order of operations is the whole point. Most prompt tools go straight from "idea" to "prompt". This one goes from "idea" to "art direction" to "prompt". That middle step is where the quality lives.
Why the human in the loop matters
The human doesn't just validate the output here. The human validates the thinking. The art direction layer this tool generates isn't a black box. You can read it, adjust it, disagree with it. And because you see the creative decisions before they become prompts, you catch a bad interpretation early, before any generation credits are spent on content that misses the mark.
At this stage of AI development, that kind of oversight makes a real difference in output quality. Models are good at pattern matching and getting better at creative interpretation, but they still benefit from a person who can say "the mood should be more restrained" or "this lighting doesn't fit the brand". That feedback, applied at the art direction level instead of the pixel level, is where you get the biggest quality gains for the least effort.
The stack
Deliberately lean:
- Framework: Next.js App Router, TypeScript, Tailwind CSS (all hand-rolled, no component libraries)
- LLM routing: OpenRouter with Gemini 2.5 Flash for vision and reference analysis, DeepSeek v3.2 for briefs and prompt generation
- Validation: Zod for all structured LLM output at runtime
- Deployment: Vercel with SSE streaming
- State: No database, no auth, no state management library. Held with the built-in React primitives. I wanted to see how minimal this could get while still being useful.
Where this is going
What I built is a tool with a human at the centre, but the architecture points toward something broader. Agentic art direction. The steps this tool performs (brief interpretation, reference analysis, visual framework, prompt writing) are all things agents will get better at on their own. Add a few more capabilities like visual trend research, output evaluation, and iterative refinement, and you have a system that can run large parts of the creative production pipeline with minimal human input.
I don't think this makes creative people irrelevant. If anything, it's the opposite. As AI takes over more of the production mechanics, human judgment moves upstream: brand strategy, creative direction, editorial taste, knowing what "good" looks like for a specific context. That's the hardest part to automate and the most valuable to get right.
The near-term future will probably look something like this. Tools that take a brand guideline, analyse current visual trends in a market, generate a visual framework, produce a first round of content, evaluate it against the brief, and iterate. All within minutes. A human checks in at key decision points rather than steering every step.
We're not there yet, but we're closer than most people think. The people producing the best AI content won't be the ones writing the best prompts by hand. They'll be the ones building the best systems around the creative process. This tool is my first step in that direction.
Try it out if you want to see how it works. Feedback is welcome, especially if you're working on similar problems.