BBVA needed a scalable, governed way to accelerate the development, validation and deployment of AI models across the Group while preserving traceability, security, risk controls and existing approval processes in a regulated financial environment.
Reported impact
BBVA accelerated AI model delivery across the Group, enabled greater team autonomy and component reuse, reduced infrastructure operating costs, and strengthened governed, auditable deployment of AI solutions in production.
BBVA implemented a new MLOps architecture with AWS inside ADA, its global cloud-based data and artificial intelligence platform, to standardize and speed the machine learning lifecycle across the bank. ADA serves as the shared platform layer for more than 6,500 users, including 1,000 data scientists, while Amazon SageMaker AI provides the cloud machine learning environment used to build, train, deploy and manage models. The architecture creates ephemeral development environments so multiple teams can experiment and validate functionality in parallel without affecting shared setups, then automatically removes resources after testing. It also automates operational tasks plus validation, traceability and control processes, and maintains a centralized audit trail for security and transparency. Approved models are integrated into BBVA’s technology operations for use cases such as personalized recommendations and financial forecasting.
Within ADA, BBVA teams create and test machine learning solutions in ephemeral Amazon SageMaker AI environments, run automated validation, traceability and control checks, record lifecycle evidence in a centralized audit trail, and then move approved models into BBVA’s production technology operations for use cases such as personalized recommendations and financial forecasting.
"BBVA has developed a new MLOps architecture together with Amazon Web Services (AWS), integrated into ADA, the bank’s global cloud-based data and artificial intelligence platform"
"The solution enables teams to work more autonomously, reuse common components and automate governance controls"
"The system automates validation, traceability and control processes to enable the safe transition of models into production"
"This transformation is based on Amazon SageMaker AI"
"creation of ephemeral development environments on AWS cloud-based machine learning infrastructure"
"the solution has reduced development times by 20 to 75 percent, and optimized infrastructure operational costs by 40 to 55 percent."
Review how this reference was scored and trace each normalized claim back to its documented source.
Scoring methodology
Customer Reference Value measures usefulness as a customer reference. Evidence Confidence measures the quality and traceability of available evidence, not independent verification.
Traceability
Field coverage shows which facts each source supports. Dates are shown only when documented by the source.
Related
AWS Sample Jul 23, 2026 BBVA As BBVA expanded the number and diversity of AI and ML solutions across the bank, it needed a more flexible, scalable, and automated MLOps architecture that improved code reuse, reproducibility, version control, and… 20% to 75% reduction in development times ADA · AWS Lambda · Amazon EventBridge AWS Sample Feb 14, 2025 MarketReader MarketReader is an artificial intelligence (AI) analytics platform providing the financial sector with data-driven explanations of real-time asset movement. During development, MarketReader experienced delays in data… 100 percent Amazon Managed Streaming for Apache Kafka · Amazon S3 · Apache Kafka Connect AWS Sample Dec 12, 2024 ElysianNxt ElysianNxt needed to replace overnight batch-style risk analysis with a cloud platform that could run large numbers of what-if analyses and regulatory stress tests quickly, while giving customers full lineage and… from 24 hours to less than 1 hour Amazon Bedrock · Amazon EKS · Amazon MSK 