Automating Risk Detection in Asset Management with Knowledge Graphs and Generative AI on AWS
Asset management is a complex field that requires portfolio managers to constantly monitor the companies in their investment universe to identify risks and opportunities that could impact their investments. While tracking direct events related to specific companies is relatively straightforward, detecting second and third-order impacts arising from events in a company’s ecosystem can be challenging.
For example, a disruption in the supply chain of a key vendor could have ripple effects on downstream manufacturers, or the loss of a top customer for a major client could pose a demand risk for its suppliers. These types of events may not always make headlines directly related to the impacted company, making them difficult to track.
In this blog post, we explore an automated solution that combines knowledge graphs and generative artificial intelligence (AI) to surface such risks by cross-referencing relationship maps with real-time news. By leveraging technologies like Amazon Neptune and Amazon Bedrock on AWS, portfolio managers can gain valuable insights into potential risks and opportunities that may not be immediately apparent.
Building the Knowledge Graph
The first step in the solution is building a knowledge graph that captures the intricate relationships between companies. Annual reports are a valuable yet often overlooked data source for knowledge graphs, as they contain accurate and reliable information about companies. By using generative AI services like Amazon Bedrock, portfolio managers can automate the process of extracting and structuring key information from annual reports to populate the knowledge graph.
Entities and relationships extracted from annual reports are carefully disambiguated to ensure accuracy before being added to the knowledge graph. Over time, this process helps build a comprehensive knowledge graph that can support valuable insights and analysis.
Processing News Articles
The next step in the solution involves automatically enriching portfolio managers’ news feeds with relevant articles that may impact their investments. Using techniques similar to processing annual reports, Amazon Bedrock is used to extract entities, attributes, and relationships from news articles, which are then linked to the knowledge graph to identify potential connections to the portfolio manager’s investments.
The final output is an enriched news feed that surfaces articles likely to impact the portfolio manager’s areas of interest and investments. By visualizing this data in a knowledge graph, portfolio managers can better understand the context and relevance of the articles.
Solution Overview
The overall architecture of the solution involves a series of steps, including uploading official reports to an S3 bucket, processing the documents using Amazon Bedrock, and enriching news articles with data from the knowledge graph. By automating these processes, portfolio managers can efficiently monitor potential risks and opportunities in real-time.
Deploying the Prototype
The prototype solution is available on GitHub and includes deployment instructions and cleanup steps for experimenting with the solution. Portfolio managers and financial professionals can leverage this prototype to explore the capabilities of knowledge graphs and generative AI in improving investment analysis and decision-making.
Summary
This blog post has demonstrated the potential of using knowledge graphs and generative AI to help portfolio managers detect second and third-order risks from news events. By combining these technologies, investment professionals can uncover valuable insights and connections that may not be immediately apparent, ultimately improving their investment analysis and decision-making processes.
If you are interested in exploring similar ideas or leveraging generative AI in your business, reach out to your AWS account manager for further exploration. The application of graph databases and AI in financial services holds great promise for augmenting investment analysis and decision-making processes, and this prototype is just the beginning of what is possible in this space.
About the Author
Xan Huang is a Senior Solutions Architect with AWS based in Singapore. He works with major financial institutions to design and build secure, scalable, and highly available solutions in the cloud. Outside of work, Xan spends most of his free time with his family and getting bossed around by his 3-year-old daughter. Connect with Xan on LinkedIn to learn more about his work and insights in the field of asset management and AI.