Cases
See Our Work in Action Through Case Studies from Leading Companies
Votorantim Cimentos faced significant challenges in converting leads into sales via WhatsApp. Additionally, there was limited understanding of which actions and initiatives were most effective on this channel. To address these issues, we implemented hGoro, an innovative recommendation engine designed to significantly boost sales and diversify product mix sold through WhatsApp.
Key Challenges:
Solution:
The hGoro recommendation engine is an AI-powered technology that continuously evaluates WhatsApp campaign performance. It predicts the most effective actions and dynamically adapts based on customer responses.
Practical Applications:
- Reactivation of inactive customers
- Hyper-personalized campaigns
- Lead prioritization (high-intent prospects)
- Automated lead nurturing & qualification
- Frequent A/B testing for continuous optimization
How It Works:
- Customer segments and available actions are registered on a web platform.
- hGoro calculates the highest-probability action for each customer group.
- Targeted campaigns are sent via WhatsApp based on these predictions.
- Customer responses (clicks/engagement) are recorded.
- hGoro learns from feedback, auto-adjusting future campaigns for maximum impact.
Results Achieved (in 5 months):
- 9.47% increase in product mix diversity
- 2x market share growth vs. pre-implementation period
- Faster lead conversion cycles
- Optimized resources & lower CAC (Customer Acquisition Cost)
hGoro has proven to be a powerful tool not only for boosting overall conversion but also for optimizing product mix and driving significant incremental revenue for Votorantim Cimentos.
This model can be effectively replicated by other businesses seeking to improve campaign performance, increase sales, and maximize ROI on WhatsApp marketing.
Nossacoop, a financial cooperative, faced a critical challenge: reactivating dormant accounts and increasing member “primary engagement” (deepening their use of the cooperative’s products/services).
To solve this, we deployed a financial product recommendation model—an AI-driven solution that delivers hyper-personalized offers with precision.
The Challenge:
The Solution:
A machine learning-powered recommendation engine that matches products to members based on:
- Product characteristics
- Member profiles
Results Achieved:
- 23% sales increase in targeted consortia campaigns (select branches)
- Higher reactivation rates of dormant accounts
- Increased product-per-member ratio ("primary engagement")
- Improved sales team productivity (data-driven prioritization)
- Enhanced member satisfaction (right offers at the right time)
This model can be adapted by other financial institutions seeking to reactivate customers, boost engagement, and drive data-led product adoption.
Direto, a financial proptech company, bridges real estate expertise with innovation and technology to deliver autonomous and efficient solutions for the credit and real estate management market.
Our Solution:
Comprehensive CRI (Real Estate Receivables Certificates) evaluationKey metrics: Volume, composition, duration, interest rates, delinquency rates, LTV, and cash flow
Automated debt acknowledgment data extractionCross-referenced with accounts receivable systems
Custom scoring model for real estate credit risk assessment
Key Outcomes:
- Enhanced portfolio management precision
- Reliable credit decision-making
- Streamlined financial analysis processes
Global Leader in Premium Tubular Solutions
Vallourec sought to enhance production quality by implementing ML-powered defect prediction and process optimization.
Our Approach:
Root cause identification for production failures
Critical variable prioritization
Defect forecasting system
Real-time process simulations
Results Achieved:
- 66.7%-83.6% defect prediction accuracy
- Optimized corrective actions through variable prioritization
The company aimed to improve welding quality in petrochemical pipeline production by predicting defects, identifying critical variables, and optimizing the process to reduce costs and increase efficiency.
Our Solution:
Defect prediction across 7 welding stages
Parameter tuning recommendations
Real-time probability scenario modeling
Operational Impact:
- 50% of defects anticipated using 3 key variables
- 10% defect reduction within first month
- Data-driven decision acceleration