Itaú Unibanco
Latin America’s largest bank, serving ~60MM individuals across corporate & investment, asset management, private banking and retail segments.
In 2022, developed and tested a quantum algorithm to predict client attrition, finding a 6% improvement in precision.
Challenges
Solution
Results
Classical ML’s Random Forrest methods are insufficient for churn prediction
Partnered with QC Ware to develop a Determinantal Point Process sampling method on a 16-qubit IBM machine and simulated it on GPUs for use today
This new model showed remarkable improvements, increasing precision by a significant 6%. A quantum-inspired version deployed today captures 2% more withdrawals.
Challenges
Classical ML’s Random Forrest methods are insufficient for churn prediction
Solution
Partnered with QC Ware to develop a Determinantal Point Process sampling method on a 16-qubit IBM machine and simulated it on GPUs for use today
Results
This new model showed remarkable improvements, increasing precision by a significant 6%. A quantum-inspired version deployed today captures 2% more withdrawals.
“We believe that the future is quantum, and it starts now with this preparation” Samurai Brito - Head of Quantum, Itaú Unibanco