Artificial Intelligence and Machine Learning - Financial Services Industry Lens

Artificial Intelligence and Machine Learning

Financial institutions have been experimenting with artificial intelligence and machine learning (AI/ML) technologies for years, but the integration of these technologies into day-to-day operations has advanced slowly due to a lack of in-house data science expertise and insufficient experience manipulating large datasets. AWS provides a set of tools that make AI/ML readily accessible to any organization. Financial institutions are using these tools to enhance customer interactions through chatbots, improve surveillance, gather trading ideas from unstructured data, and customize product offerings, among many other use cases.

Financial services AI/ML architectures supporting these use cases share the following characteristics:

  • They have a secure architecture to protect code and model artifacts.

  • They have self-service capabilities for model development and training environments with pre-defined security configurations.

  • They use a CI/CD pipeline integrated with change control systems for model deployment.

  • They automate end to end evidence capture of the entire model development lifecycle across development, training, and deployment.

Reference Architecture

Figure 3: Reference architecture for an AI/ML pipeline