People, Prompts & Pixels: Kulfi Collective’s Blueprint for Personalized Content at Scale
by Nikita D’souza, Assistant VP of Marketing @ Kulfi Collective
If you’ve ever opened Netflix and thought, “How do they know exactly what I want to watch tonight?” or scrolled Instagram and found a meme that felt like it was made just for you—you’ve felt the shift from scale to personalization at scale.
Personalization is no longer a luxury. It’s the expectation.
And as AI reshapes how we imagine, make, and scale content, that belief matters more than ever. The internet doesn’t just want more content—it wants content that feels contextual and culturally resonant. Fast. At scale. And without losing the craft and nuance that makes it human.
For us at Kulfi Collective, the answer to creating that content lies in a deceptively simple formula: people, prompts, and pixels.
People bring cultural judgment, creative instinct, and taste.
Prompts are the new art direction—the bridge between human intent and machine generation.
Pixels are the AI-powered assets, refined with craft to make them brand-ready.
We don’t see AI as replacing creativity. We see it as rewiring creativity for speed, scale, and personalization. And when done right, it doesn’t erase the human touch, it amplifies it.
Reimagining the Creative Assembly Line
The internet has reshaped how fast stories need to be told. So, at Kulfi, we re-engineered our workflows into hybrid pipelines where humans and AI collaborate.
Think of it as an assembly line that starts with data, translates briefs into prompts, and generates infinite variations. But here’s the catch: speed without judgment is chaos. That’s why we built Kulfi OS, a new way to manage, govern and approve assets at scale, ensuring brand fidelity, cultural resonance, and creative quality before anything sees the light of day.
As Jayson D’Mello, Chief Innovations Officer at Kulfi, says:
“While AI gives us incredible velocity and scale, velocity without control is just a faster way to damage your brand. Our philosophy is simple: every AI-driven workflow needs a parallel system of brand governance. That’s the ‘Brand Brain’ for that creative engine, ensuring every single asset is perfectly on-brand. This human plus machine guardrail keeps scale from becoming soulless."
This balance of rigor and play is what makes the Kulfi approach unique. And you can see it in action in two projects: Agoda and Birla Opus.
Case 1: Agoda — 25,000 Holidays, One Perfect Trip for You
Travel is a joy. But deciding where to go, what to see, what to do? Overwhelming! Digital travel platform Agoda wanted to fix that, and turn it into a shareable brand moment.
Our solution: a personalized content engine that transformed indecision into storytelling.
First, we designed a microsite with a playful tarot-style quiz to capture user preferences. Behind the scenes, we mapped 25,000 possible travel archetypes by mixing budgets, companions, and experiences. An LLM-powered reasoning layer scored and picked the one perfect trip.
Google Imagen 3 generated bespoke visuals of each destination, while a modular shoot with Bollywood star, Ayushmann Khurrana, provided templated lines stitched together with flawless AI-powered lip-sync. Within minutes, each traveler received a personalized video with Ayushmann greeting them by name, revealing their destination, and showcasing highlights. Plus a complete, detailed day-wise itinerary!
As director Mohit Bhasin put it: “This wasn’t just another celebrity shoot—it was a film engineered for exponential scale. Every line, pause, and camera block had to anticipate where AI would slot in variables without breaking natural flow. We had to honor both the creative vision and the technical constraints, while still keeping the film cinematic.”
The outcome: 30,000+ personalized videos generated, millions of interactions, and proof that hyper-personalization can feel magical, not mechanical.
Case 2: Birla Opus — Painting the Internet With 2,300 Colors
When Birla Opus entered India’s crowded paint market, they didn’t just need to showcase 2,300 colors. They needed each one to feel like a mood, a story, an invitation to imagine.
Traditionally, producing this library would take months of manual shoots and ballooning costs. We engineered a different path: a bespoke AI-assisted system that combined color psychology, scalable prompts, workflow automation and artist-in-the-loop refinement.
Our creative directors first mapped the emotions behind each color family, building a design system of textures, objects, and lighting. Next, we fine-tuned LLMs to generate variations. Our custom-built bot, Kulfi Automate, interfaced with generative tools to output thousands of images, and our artists polished every shade for accuracy and resonance.
As Creative Director, Sushrita Bhattacharjee, explained:
“Scalability comes from doing the hard, manual work first. We tested and tweaked variables shade by shade until we found the right balance—light, medium, dark—that could capture the mood of a color family. Once the system worked on 100 shades, scaling it 2,000 times was effortless.”
The result: 2,300+ color-perfect images delivered 75% faster, at 95% lower cost, with a 2x faster feedback process. More importantly, each shade carried emotional resonance, not just a hex code.
A Glimpse into the Future
From travel itineraries to paint palettes, audiences today expect stories that feel like they were made just for them. The challenge (and the opportunity!) is delivering at the speed of culture, without losing what makes creativity human.
At Kulfi, our blueprint is simple: people, prompts, pixels.
That’s how we’re building creative systems that are fast enough for the internet, flexible enough for brands, and still deeply personal.
Because in the end, the internet doesn’t remember ads.
It remembers stories that feel like they were made for you.
Nikita D’souza, Kulfi Collective
