A marketing agency needed a faster way to turn project outcomes into branded case studies. After manually writing dozens by hand, I built an automated storytelling system that generates three assets — a one-pager, a social carousel, and a blog post — from one short form.
The Problem
Every agency faces the same bottleneck:
you deliver great client work, but the proof gets buried in Google Docs or inboxes.
Each case study followed the same structure — company, service, story, results — yet every time, the process reset.
The copy team would rewrite sections, the design team would rebuild layouts, and publishing would lag behind.
Even with templates, brand tone and format drifted between versions. The process wasn’t broken — it was just too manual to scale.
They needed a way to systemize storytelling without losing the creativity and human voice that made each project unique.
The Approach
I didn’t start with automation.
I started with repetition.
For two months, I wrote and designed case studies manually.
That was the key.
The repetition revealed what stayed consistent — and what truly needed human judgment.
Once I saw the pattern, I mapped the process end-to-end and rebuilt it as a system.
The Solution
The new workflow turns a 2–3 week production cycle into a 6-minute automation powered by Tally, GPT, Make (Integromat), and Canva.
- Collect the essentials with Tally
- Company
- Service
- Story / Information
- Results
- Translate data to structured copy with GPT
- Automate design with Canva Autofill
- Extend to blog version for content reuse
- Run through 22 operations
A simple intake form captures four variables:
Using a JSON prompt, GPT generates text blocks that mirror the agency’s brand voice — title, hook, summary, and paragraph structure.
The JSON format acts as a template for tone and rhythm, ensuring consistency across every output.
Each section of the JSON maps to a Canva template.
The system populates the one-pager and carousel instantly, using pre-defined styles, brand fonts, and colors.
The same data feeds into a long-form blog layout in Google Docs — ready for minor edits and SEO tagging.
In Make, these modules handle formatting, duplication, QA steps, and file delivery — finishing the loop in under 6 minutes.
Do it manually until you can see the pattern. Then systemize it.
That principle shaped every part of this build.
The Result
The agency now has a “case study engine” —
a system that can generate three cohesive assets from a single intake.
- Production time: 2–3 weeks → 1 day
- Output consistency: 100% tone-aligned, 0% drift
- Assets per project: One-pager, carousel, and blog, all synchronized
- Human focus: Writers now refine insights instead of rebuilding structure
Every client story now moves from invisible → visible → shareable — without manual repetition.
Behind the Build
This wasn’t just about creating templates or AI text.
It was about defining the invisible structure of storytelling and turning it into a system anyone could use.
Automation doesn’t replace human creativity — it protects it.
It removes the repetitive steps so the real thinking can stay human.
Conclusion
By turning a creative process into a structured automation, we built a repeatable storytelling system that scales without sacrificing tone, design, or brand voice.
A quiet backend that lets the story speak louder.
Credits
- Built by: Harrison Caldeira
- Client: Confidential
- Tools: Tally, GPT, Canva, Make (Integromat)
- Time to Build: 2 months of manual testing → 1 day automated system