Today, we are pleased to announce a new integration between Amazon Quick Research and S&P Global. This integration brings both S&P Global Energy news, research and insights and S&P Global Market Intelligence data into one in-depth research agent for Accelerated Research clients.
S&P Global integration expands the capabilities of Instant Research so that business professionals can analyze multiple data sources, including global energy news and premium financial intelligence, in a single workspace, eliminating the need to switch between platforms and turning weeks of research into minutes of focused insight generation. Quick Suite connects information through internal repositories, popular applications, AWS services, and Model Reference Protocol (MCP) integration, for over 1,000 apps. This agentic AI application is reshaping the way teams work by transforming the way they discover insights, conduct deep research, automate tasks, visualize data, and take action across apps.
In this post, we explore the solution architecture of S&P Global’s data sets and integration with Instant Research.
solution overview
S&P Global Pioneered two MCP server implementations on AWS so organizations can easily integrate trusted financial services and energy content into AI-powered workflows while maintaining the quality, security, and reliability business leaders demand.
“Our collaboration with AWS expands how S&P Global delivers trusted intelligence through next-generation agentic AI experiences. By working with leading AI companies, our goal is to ensure clients can access our trusted data and insights wherever their workflow occurs.”
– Bhavesh Dayalji, Chief AI Officer of S&P Global and CEO of Kensho.
S&P Global Energy: Comprehensive Commodity and Energy Intelligence
S&P Global Energy Integration, now available in Amazon Quick Research, uses AI ready data MCP Server will provide comprehensive access to commodity and energy market information across global markets spanning the oil, gas, power, metals, clean energy, agriculture and shipping sectors. Building on S&P Global’s reputation as a trusted markets authority, MCP Server utilizes hundreds of thousands of expert-crafted documents, including analyses, commentaries and news articles reflecting decades of industry expertise.
The solution provides a unique multi-horizon perspective, ranging from daily market updates to one-year outlooks and 20+ year scenario analyses. With data refreshed every 30 minutes, business leaders get real-time access to commodity and energy intelligence, dramatically accelerating the speed of decisions when discovering regulatory challenges, investment opportunities or environmental implications.
S&P Global Market Intelligence: reliable financial intelligence
The S&P Global Market Intelligence integration, now available in Amazon Quick Research, uses the Kensho LLM-ready API MCP Server developed by Kensho, S&P Global’s AI innovation hub. This MCP server integrates seamlessly with Amazon Quick Research to make trusted financial data accessible through natural language queries. Financial professionals can access S&P Capital IQ financials, earnings call transcripts, company information, transactions and more just by asking questions.
The Kensho solution addresses a critical challenge in financial services: making vast stores of financial data instantly accessible without the need for complex query languages ​​or technical expertise. Engineering, product, and business teams can save significant time and resources by converting the hours required for data extraction into conversational queries that return accurate, reliable information in seconds.
solution architecture
S&P Global’s MCP server architecture is shown in the following figure. When using one of the S&P integrations, traffic flows from Quick Research through Amazon API Gateway to an AWS Application Load Balancer with MCP services hosted on Amazon Elastic Kubernetes Service (Amazon EKS). The MCP server uses data hosted in Amazon S3 and AWS Relational Database Service for PostgreSQL for structured data, and Amazon OpenSearch service for vector storage. This architecture provides enterprise-ready MCP servers with deep security, automatic scaling, and broad observability.
mcp is an open standard that supports seamless communication between AI agents and external data sources, devices, and services. MCP operates on a client-server architecture where MCP servers handle tool calls, typically involving multiple API calls, and expose business logic implementations as callable functions. It enables AI agents to dynamically discover capabilities, interact across features, and securely share context, all critical requirements for enterprise-grade applications.
S&P Global’s solution consists of the following key building blocks:
- Automated Data Pipelines with Amazon Bedrock: At the heart of the solution is a Retrieval Augmented Generation (RAG) data ingestion pipeline using Amazon Bedrock. This pipeline transforms raw market data AI ready dataDocuments from S&P Global’s proprietary repository undergo preprocessing, chunking, and enrichment before being converted to vector embeddings using the Bedrock hosted Cohair embedding model, The ingestion pipeline runs on a scheduled basis, refreshing the OpenSearch vector store every 30 minutes for real-time access to energy data,
- Vector and semantic search: Amazon OpenSearch serves as the vector database, storing the embeddings generated by Bedrock and enabling semantic search capabilities in S&P Global’s energy data. The OpenSearch Vector Store is optimized for high-dimensional vector operations, supporting fast similarity searches that power the MCP Server’s ability to find contextually relevant information in response to natural language queries.
- Flexibility and scale: This solution uses Amazon EKS to host all MCP server solutions with two production clusters enabling traffic segmentation and failover capabilities. This dual-cluster approach provides continuous availability even during unexpected failures. Both the Cluster Autoscaler and the Horizontal Pod Autoscaler enable dynamic scaling based on demand. MCP servers are built with the FastMCP framework, providing high-performance HTTP endpoints that comply with the Streamable HTTP Transport specification required by the MCP protocol.
- Security: Security is built into every layer of the solution. The API Gateway serves as the endpoint for MCP server access. S&P Global’s enterprise identity provider is used OAuth Certification. API Gateway is secured with AWS Web Application Firewall (WAF) with advanced threat detection. AWS IAM roles and policies enforce least privilege principles, so that each component has only the permissions it needs. AWS Secrets Manager securely stores credentials to access resources and AWS services. AWS Security Groups and VPC configurations provide network isolation, while TLS 1.2+ with AWS Certificate Manager verifies all data in transit and remains encrypted. This multi-layered security includes deep security controls.
- Observability: Amazon CloudWatch provides centralized logging, metrics collection, and real-time monitoring of the entire pipeline from data ingestion through MCP server responses. AWS CloudTrail captures detailed API activity logs and audit trails, which are essential for compliance in regulated industries.
conclusion
Together, built on AWS and integrated into Amazon Quick Research, these MCP servers demonstrate S&P Global’s vision for the future of financial services and energy intelligence: maintaining the trust, accuracy, and depth that business leaders need while embracing the transformative potential of AI to make that intelligence more accessible, actionable, and integrated into modern workflows.
Ready to get started? Please see Quick Research third party data for more information.
About the authors
john einkoff is a product leader at Seattle-based AWS, where he focuses on building AI-powered tools that help businesses synthesize information and accelerate research. With over a decade of experience across digital health, cloud computing, and AI products at Amazon, he has led cross-functional teams in product management, engineering, and design to deliver innovative solutions for customers around the world.
Prashant Ponnoth is an AWS solutions architect supporting global financial services with over 20 years of industry and technology experience with cloud migration, modernization, and building large-scale distributed systems. His areas of interest are machine learning, containers/Kubernetes, and open-source technologies. At AWS, he is part of the Machine Learning technical area community and focusing on Amazon Bedrock, Amazon SageMaker AI, Amazon Bedrock AgentCore services.
Brandon Pominville is a senior solutions architect at AWS, based in New York, where he works with global financial services customers to build secure, scalable data and AI platforms in the cloud. With over 20 years of experience in financial services, enterprise data platforms, and cloud computing, he specializes in transforming business needs into technology solutions. Outside of work, Brandon enjoys spending time with his family outdoors or on a cruise ship and playing volleyball.
