The network for creativity
Join 1.25M professional creatives like you
Connect with clients, get discovered, and run your business 100% commission-free
Creatives on Contra have earned over $150M and we are just getting started
Back to feedPost
Case study: An AI content pipeline we built for a media client.
The challenge: turn a stream of source documents (including scanned PDFs) into automated daily email digests.
What made it hard:
→ Half the inputs were scanned PDFs of variable quality. Vision models for OCR > legacy OCR engines.
→ Hundreds of inputs/day. Pipeline extracts, groups, dedupes, rewrites, and scores by importance.
→ LLM landscape moves fast — we built a swappable provider layer.
Stack: Next.js + Python + provider-agnostic model layer.
If you're staring at messy inputs and wondering whether an LLM can make sense of them: yes — but the engineering matters more than the model.
The network for creativity
Join 1.25M professional creatives like you
Connect with clients, get discovered, and run your business 100% commission-free
Creatives on Contra have earned over $150M and we are just getting started
Related posts
Hot take: The more artificial the technology gets, the more human they make the branding become.
Look at Granola, Wispr Flow and of course Anthropic, and dozens of newer AI startups. Serif typography is everywhere.
For years, tech chased clean sans serifs, gradients, efficiency, and neutrality. Now the trend is warmth, personality, and a more human feel.
Maybe that's the irony of AI.
The branding is getting more human because the technology is getting less human. Feels like a natural correction.
I joined the Config Makeathon and built BriefOS — a client brief operating system for designers, studios, and freelancers.
The problem I wanted to solve is something almost every designer deals with:
Client projects rarely start clean.
They usually start with scattered notes, vague goals, random references, unclear positioning, mixed opinions, and no approved direction. That chaos creates slow starts, endless revisions, weak strategy, and confusion between the designer and the client.
So I built BriefOS to turn that mess into a clear, structured website blueprint.
The idea:
Client chaos → AI signal extraction → website blueprint → visual direction → wireframe → client approval
Inside BriefOS, a designer can collect messy client input, organize it into useful categories, extract important signals, build a sitemap, generate copy directions, choose visual routes, adjust wireframes, and send decisions into a client approval room.
I wanted the product to feel less like a generic AI tool and more like a real project command center.
What I built:
Command Center to track the full project workflow
Intake Wall for messy client notes, links, screenshots, and references
Signal Extractor to identify offer, audience, CTA, tone, gaps, and contradictions
Blueprint Builder for sitemap, homepage structure, copy, and conversion path
Visual Lab for mood, palette, typography, layout direction, and style routes
Wireframe Studio for reviewing and refining the homepage structure
Approval Room for client decisions, comments, approvals, and revision requests
I also focused on making the experience interactive instead of just “AI magically does everything.”
The user can tag notes, approve or reject signals, choose direction, reorder sections, compare options, adjust wireframes, and approve final decisions.
Built using Figma Make as part of the Config Makeathon.
The goal was simple:
Help designers move from confusion to clarity faster — while keeping strategy, copy, design, and client approval connected in one workflow.
Still more I’d love to improve:
deeper real data inputs
stronger drag-and-drop behavior
better client presentation mode
more polished animations
live export options
more visual direction templates
But for this version, I’m happy with the core concept: a useful AI workflow product built around a real design problem.
preview link:https://www.figma.com/make/LhthPqSGpwgLBKVoJdQLHG/Begin-Project?fullscreen=1&t=aPQZNZHYQiD53qn0-1&code-node-id=0-9
#ConfigMakeathon @figma #FigmaMake #ProductDesign #UIDesign #AIDesign #WebDesign #DesignTools #Contra
This is actually a really useful idea. Turning messy client briefs into a clear workflow solves a real problem.
7 DAYS LEFT ⏳
Jumped into the Figma’s Config Makeathon to finally build the idea that's been living rent-free in my head:
A community app where AI transforms tired public spaces into before/after visions. Vote, remix, repeat.
First design's done. What do you think?
ok this is looking so sharp 👀
Challenges
View all
Trending
Claude
Claude has entered the design space. How are you using Claude Design?
Contra University
Learn from expert creatives how to earn more using next-gen AI tools.
MagicPath
The canvas is infinite, and exploration is becoming the workflow. How are you using MagicPath?
creativeaiflow
Creative AI workflows are evolving. What tools do you use, and what are their strengths and weaknesses?
freelancerlife
Freelancer life is wins, pivots, and everything in between. What’s yours right now?