Company
Dicer.ai is a performance marketing SaaS platform that combines AI with marketing expertise to optimize ad creatives and campaign strategies for higher conversion rates. The platform connects creative development, performance analytics, and media buying into a unified solution.
The Challenge
Marketing teams face a critical dilemma with AI-generated content: while it promises speed and personalization, 71% of marketers report that AI tools produce generic or bland creative1. Traditional video ad production requires days of manual work per asset, limiting the ability to test variations and respond to market changes. Dicer.ai set out to automate ad creation while maintaining quality and brand authenticity—but generating high-quality ads at scale wasn't enough. They also needed to forecast performance before deployment, because without predicting which ads would actually perform, marketing teams would just be producing mediocre content faster.
The Solution
Dicer.ai partnered with A.Team to embed an AI Engineer with expertise in visual AI systems—a less mature domain than text-focused LLMs. Rather than building another generic video generator, the A.Teamer created an end-to-end competitive intelligence pipeline. The system maps each client's market landscape, scraping the web to understand positioning and identify competitors, then analyzes their top-performing ads scene-by-scene to extract creative patterns driving success. Using these insights, it generates multiple video variations that combine proven market formulas with client brand assets—reverse-engineering what works while maintaining authenticity. The breakthrough was integrating performance forecasting to predict ad reach before deployment, addressing the ROI uncertainty that makes most AI-generated content an expensive gamble.
Brafton 1
Technologies Used:
Python: Built data pipelines and integrated AI models for rapid iteration
Google Gemini: Analyzed competitor videos to extract scene-by-scene storyboards and creative patterns
Semantic Search: Tied brand assets to high-performing creative contexts from historical data
LLMs: Generated ad variations based on competitive intelligence insights
Data Processing Pipelines: Scraped and analyzed web data to map competitive landscapes
"It's easy to generate ads, but measuring how successful they will be is another challenge entirely. Our ability to forecast ad reach was just as important as the generation itself.”

— A.Team AI Architect
The Results
Dicer.ai's AI-powered system turned their clients' creative production from a multi-day manual process into an automated pipeline that delivered multi-variate video ads in under two hours.
Production Speed
Reduced ad creation from days to 1-2 hours, enabling rapid market response
ROI Forecasting
Predicted reach pre-launch to focus spend on top-performing ads
Testing Scale
Produced 5 video variations simultaneously for multi-variant performance testing
Competitive Edge
Extracted winning patterns from competitor ads while maintaining brand authenticity